natsort: Simple yet flexible natural sorting in Python.

• Source Code: https://github.com/SethMMorton/natsort
• Downloads: https://pypi.org/project/natsort/
• Documentation: https://natsort.readthedocs.io/

Please see the GitHub main page for everything else, including

• Quick description
• Basic examples
• FAQ
• Requirements and optional dependencies
• Installation instructions
• Testing instructions
• Deprecation schedule

How Does Natsort Work?

natsort works by breaking strings into smaller sub-components (numbers or everything else), and returning these components in a tuple. Sorting tuples in Python is well-defined, and this fact is used to sort the input strings properly. But how does one break a string into sub-components? And what does one do to those components once they are split? Below I will explain the algorithm that was chosen for the natsort module, and some of the thinking that went into those design decisions. I will also mention some of the stumbling blocks I ran into because getting sorting right is surprisingly hard.

If you are impatient, you can skip to TL;DR 1 - The Simple “No Special Cases” Algorithm for the algorithm in the simplest case, and TL;DR 2 - Handling Crappy, Real-World Input to see what extra code is needed to handle special cases.

First, How Does Natural Sorting Work At a High Level?

If I want to compare ‘2 ft 7 in’ to ‘2 ft 11 in’, I might do the following

>>> '2 ft 7 in' < '2 ft 11 in'
False

We as humans know that the above should be true, but why does Python think it is false? Here is how it is performing the comparison:

'2' <=> '2' ==> equal, so keep going
' ' <=> ' ' ==> equal, so keep going
'f' <=> 'f' ==> equal, so keep going
't' <=> 't' ==> equal, so keep going
' ' <=> ' ' ==> equal, so keep going
'7' <=> '1' ==> different, use result of '7' < '1'

‘7’ evaluates as greater than ‘1’ so the statement is false. When sorting, if a value is less than another it is placed first, so in our above example ‘2 ft 11 in’ would end up before ‘2 ft 7 in’, which is not correct. What to do?

The best way to handle this is to break the string into sub-components of numbers and non-numbers, and then convert the numeric parts into float() or int() types. This will force Python to actually understand the context of what it is sorting and then “do the right thing.” Luckily, it handles sorting lists of strings right out-of-the-box, so the only hard part is actually making this string-to-list transformation and then Python will handle the rest.

'2 ft 7 in'  ==> (2, ' ft ', 7,  ' in')
'2 ft 11 in' ==> (2, ' ft ', 11, ' in')

When Python compares the two, it roughly follows the below logic:

2       <=> 2      ==> equal, so keep going
' ft '  <=> ' ft ' ==> a string is a special type of sequence - evaluate each character individually
                   ||
                   -->
                      ' ' <=> ' ' ==> equal, so keep going
                      'f' <=> 'f' ==> equal, so keep going
                      't' <=> 't' ==> equal, so keep going
                      ' ' <=> ' ' ==> equal, so keep going
                  <== Back to parent sequence
7 <=> 11 ==> different, use the result of 7 < 11

Clearly, seven is less than eleven, so our comparison is as we expect, and we would get the sorting order we wanted.

At its heart, natsort is simply a tool to break strings into tuples, turning numbers in strings (i.e. '79') into ints and floats as it does this.

Natsort’s Approach

Decomposing Strings Into Sub-Components

The first major hurtle to overcome is to decompose the string into sub-components. Remarkably, this turns out to be the easy part, owing mostly to Python’s easy access to regular expressions. Breaking an arbitrary string based on a pattern is pretty straightforward.

>>> import re
>>> re.split(r'(\d+)', '2 ft 11 in')
['', '2', ' ft ', '11', ' in']

Clear (assuming you can read regular expressions) and concise.

The reason I began developing natsort in the first place was because I needed to handle the natural sorting of strings containing real numbers, not just unsigned integers as the above example contains. By real numbers, I mean those like -45.4920E-23. natsort can handle just about any number definition; to that end, here are all the regular expressions used in natsort:

>>> unsigned_int               = r'([0-9]+)'
>>> signed_int                 = r'([-+]?[0-9]+)'
>>> unsigned_float             = r'((?:[0-9]+\.?[0-9]*|\.[0-9]+)(?:[eE][-+]?[0-9]+)?)'
>>> signed_float               = r'([-+]?(?:[0-9]+\.?[0-9]*|\.[0-9]+)(?:[eE][-+]?[0-9]+)?)'
>>> unsigned_float_no_exponent = r'((?:[0-9]+\.?[0-9]*|\.[0-9]+))'
>>> signed_float_no_exponent   = r'([-+]?(?:[0-9]+\.?[0-9]*|\.[0-9]+))'

Note that "inf" and "nan" are deliberately omitted from the float definition because you wouldn’t want (for example) "banana" to be converted into ['ba', 'nan', 'a'], Let’s see an example:

>>> re.split(signed_float, 'The mass of 3 electrons is 2.732815068E-30 kg')
['The mass of ', '3', ' electrons is ', '2.732815068E-30', ' kg']

Note

It is a bit of a lie to say the above are the complete regular expressions. In the actual code there is also handling for non-ASCII unicode characters (such as ⑦), but I will ignore that aspect of natsort in this discussion.

Now, when the user wants to change the definition of a number, it is as easy as changing the pattern supplied to the regular expression engine.

Choosing the right default is hard, though (well, in this case it shouldn’t have been but I was rather thick-headed). In retrospect, it should have been obvious that since essentially all the code examples I had/have seen for natural sorting were for unsigned integers, I should have made the default definition of a number an unsigned integer. But, in the brash days of my youth I assumed that since my use case was real numbers, everyone else would be happier sorting by real numbers; so, I made the default definition of a number a signed float with exponent. This astonished a lot of people (and some people aren’t very nice when they are astonished). Starting with natsort version 4.0.0 the default number definition was changed to an unsigned integer which satisfies the “least astonishment” principle, and I have not heard a complaint since.

Coercing Strings Containing Numbers Into Numbers

There has been some debate on Stack Overflow as to what method is best to coerce a string to a number if it can be coerced, and leaving it alone otherwise (see this one for coercion and this one for checking for some high traffic questions), but it mostly boils down to two different solutions, shown here:

>>> def coerce_try_except(x):
...     try:
...         return int(x)
...     except ValueError:
...         return x
...
>>> def coerce_regex(x):
...     # Note that precompiling the regex is more performant,
...     # but I do not show that here for clarity's sake.
...     return int(x) if re.match(r'[-+]?\d+$', x) else x
...

Here are some timing results run on my machine:

In [0]: numbers = list(map(str, range(100)))  # A list of numbers as strings

In [1]: not_numbers = ['banana' + x for x in numbers]

In [2]: %timeit [coerce_try_except(x) for x in numbers]
10000 loops, best of 3: 51.1 µs per loop

In [3]: %timeit [coerce_try_except(x) for x in not_numbers]
1000 loops, best of 3: 289 µs per loop

In [4]: %timeit [coerce_regex(x) for x in not_numbers]
10000 loops, best of 3: 67.6 µs per loop

In [5]: %timeit [coerce_regex(x) for x in numbers]
10000 loops, best of 3: 123 µs per loop

What can we learn from this? The try: except method (arguably the most “pythonic” of the solutions) is best for numeric input, but performs over 5X slower for non-numeric input. Conversely, the regular expression method, though slower than try: except for both input types, is more efficient for non-numeric input than for input that can be converted to an int. Further, even though the regular expression method is slower for both input types, it is always at least twice as fast as the worst case for the try: except.

Why do I care? Shouldn’t I just pick a method and not worry about it? Probably. However, I am very conscious about the performance of natsort, and want it to be a true drop-in replacement for sorted() without having to incur a performance penalty. For the purposes of natsort, there is no clear winner between the two algorithms - the data being passed to this function will likely be a mix of numeric and non-numeric string content. Do I use the try: except method and hope the speed gains on numbers will offset the non-number performance, or do I use regular expressions and take the more stable performance?

It turns out that within the context of natsort, some assumptions can be made that make a hybrid approach attractive. Because all strings are pre-split into numeric and non-numeric content before being passed to this coercion function, the assumption can be made that if a string begins with a digit or a sign, it can be coerced into a number.

>>> def coerce_to_int(x):
...     if x[0] in '0123456789+-':
...         try:
...             return int(x)
...         except ValueError:
...             return x
...     else:
...         return x
...

So how does this perform compared to the standard coercion methods?

In [6]: %timeit [coerce_to_int(x) for x in numbers]
10000 loops, best of 3: 71.6 µs per loop

In [7]: %timeit [coerce_to_int(x) for x in not_numbers]
10000 loops, best of 3: 26.4 µs per loop

The hybrid method eliminates most of the time wasted on numbers checking that it is in fact a number before passing to int(), and eliminates the time wasted in the exception stack for input that is not a number.

That’s as fast as we can get, right? In pure Python, probably. At least, it’s close. But because I am crazy and a glutton for punishment, I decided to see if I could get any faster writing a C extension. It’s called fastnumbers and contains a C implementation of the above coercion functions called fast_int(). How does it fair? Pretty well.

In [8]: %timeit [fast_int(x) for x in numbers]
10000 loops, best of 3: 30.9 µs per loop

In [9]: %timeit [fast_int(x) for x in not_numbers]
10000 loops, best of 3: 30 µs per loop

During development of natsort, I wanted to ensure that using it did not get in the way of a user’s program by introducing a performance penalty to their code. To that end, I do not feel like my adventures down the rabbit hole of optimization of coercion functions was a waste; I can confidently look users in the eye and say I considered every option in ensuring natsort is as efficient as possible. This is why if fastnumbers is installed it will be used for this step, and otherwise the hybrid method will be used.

Note

Modifying the hybrid coercion function for floats is straightforward.

>>> def coerce_to_float(x):
...     if x[0] in '.0123456789+-' or x.lower().lstrip()[:3] in ('nan', 'inf'):
...         try:
...             return float(x)
...         except ValueError:
...             return x
...     else:
...         return x
...

TL;DR 1 - The Simple “No Special Cases” Algorithm

At this point, our natsort algorithm is essentially the following:

>>> import re
>>> def natsort_key(x, as_float=False, signed=False):
...     if as_float:
...         regex = signed_float if signed else unsigned_float
...     else:
...         regex = signed_int if signed else unsigned_int
...     split_input = re.split(regex, x)
...     split_input = filter(None, split_input)  # removes null strings
...     coerce = coerce_to_float if as_float else coerce_to_int
...     return tuple(coerce(s) for s in split_input)
...

I have written the above for clarity and not performance. This pretty much matches most natural sort solutions for python on Stack Overflow (except the above includes customization of the definition of a number).

Special Cases Everywhere!

If what I described in TL;DR 1 were all that natsort needed to do then there probably wouldn’t be much need for a third-party module, right? Probably. But it turns out that in real-world data there are a lot of special cases that need to be handled, and in true 80%/20% fashion, the majority of the code in natsort is devoted to handling special cases like those described below.

Sorting Filesystem Paths

The first major special case I encountered was sorting filesystem paths (if you go to the link, you will see I didn’t handle it well for a year… this was before I fully realized how much functionality I could really add to natsort). Let’s apply the natsort_key() from above to some filesystem paths that you might see being auto-generated from your operating system:

>>> paths = ['Folder (10)/file.tar.gz',
...          'Folder/file.tar.gz',
...          'Folder (1)/file (1).tar.gz',
...          'Folder (1)/file.tar.gz']
>>> sorted(paths, key=natsort_key)
['Folder (1)/file (1).tar.gz', 'Folder (1)/file.tar.gz', 'Folder (10)/file.tar.gz', 'Folder/file.tar.gz']

Well that’s not right! What is 'Folder/file.tar.gz' doing at the end? It has to do with the numerical ASCII code assigned to the space and / characters in the ASCII table. According to the ASCII table, the space character (number 32) comes before the / character (number 47). If we remove the common prefix in all of the above strings ('Folder'), we can see why this happens:

>>> ' (1)/file.tar.gz' < '/file.tar.gz'
True
>>> ' ' < '/'
True

This isn’t very convenient… how do we solve it? We can split the path across the path separators and then sort. A convenient way do to this is with the Path.parts property from pathlib:

>>> import pathlib
>>> sorted(paths, key=lambda x: tuple(natsort_key(s) for s in pathlib.Path(x).parts))
['Folder/file.tar.gz', 'Folder (1)/file (1).tar.gz', 'Folder (1)/file.tar.gz', 'Folder (10)/file.tar.gz']

Almost! It seems like there is some funny business going on in the final filename component as well. We can solve that nicely and quickly with Path.suffixes and Path.stem.

>>> def decompose_path_into_components(x):
...     path_split = list(pathlib.Path(x).parts)
...     # Remove the final filename component from the path.
...     final_component = pathlib.Path(path_split.pop())
...     # Split off all the extensions.
...     suffixes = final_component.suffixes
...     stem = final_component.name.replace(''.join(suffixes), '')
...     # Remove the '.' prefix of each extension, and make that
...     # final component a list of the stem and each suffix.
...     final_component = [stem] + [x[1:] for x in suffixes]
...     # Replace the split final filename component.
...     path_split.extend(final_component)
...     return path_split
...
>>> def natsort_key_with_path_support(x):
...     return tuple(natsort_key(s) for s in decompose_path_into_components(x))
...
>>> sorted(paths, key=natsort_key_with_path_support)
['Folder/file.tar.gz', 'Folder (1)/file.tar.gz', 'Folder (1)/file (1).tar.gz', 'Folder (10)/file.tar.gz']

This works because in addition to breaking the input by path separators, the final filename component is separated from its extensions as well. Then, each of these separated components is sent to the natsort algorithm, so the result is a tuple of tuples. Once that is done, we can see how comparisons can be done in the expected manner.

>>> a = natsort_key_with_path_support('Folder (1)/file (1).tar.gz')
>>> a
(('Folder (', 1, ')'), ('file (', 1, ')'), ('tar',), ('gz',))
>>>
>>> b = natsort_key_with_path_support('Folder/file.tar.gz')
>>> b
(('Folder',), ('file',), ('tar',), ('gz',))
>>>
>>> a > b
True

Comparing Different Types on Python 3

The second major special case I encountered was sorting of different types. If you are on Python 2 (i.e. legacy Python), this mostly doesn’t matter too much since it uses an arbitrary heuristic to allow traditionally un-comparable types to be compared (such as comparing 'a' to 1). However, on Python 3 (i.e. Python) it simply won’t let you perform such nonsense, raising a TypeError instead.

You can imagine that a module that breaks strings into tuples of numbers and strings is walking a dangerous line if it does not have special handling for comparing numbers and strings. My imagination was not so great at first. Let’s take a look at all the ways this can fail with real-world data.

>>> def natsort_key_with_poor_real_number_support(x):
...     split_input = re.split(signed_float, x)
...     split_input = filter(None, split_input)  # removes null strings
...     return tuple(coerce_to_float(s) for s in split_input)
>>>
>>> sorted([5, '4'], key=natsort_key_with_poor_real_number_support)
Traceback (most recent call last):
    ...
TypeError: ...
>>>
>>> sorted(['12 apples', 'apples'], key=natsort_key_with_poor_real_number_support)
Traceback (most recent call last):
    ...
TypeError: ...
>>>
>>> sorted(['version5.3.0', 'version5.3rc1'], key=natsort_key_with_poor_real_number_support)
Traceback (most recent call last):
    ...
TypeError: ...

Let’s break these down.

1. The integer 5 is sent to re.split which expects only strings or bytes, which is a no-no.
2. natsort_key_with_poor_real_number_support('12 apples') < natsort_key_with_poor_real_number_support('apples') is the same as (12.0, ' apples') < ('apples',), and thus a number gets compared to a string [1] which also is a no-no.
3. This one scores big on the astonishment scale, especially if one accidentally uses signed integers or real numbers when they mean to use unsigned integers. natsort_key_with_poor_real_number_support('version5.3.0') < natsort_key_with_poor_real_number_support('version5.3rc1') is the same as ('version', 5.3, 0.0) < ('version', 5.3, 'rc', 1.0), so in the third element a number gets compared to a string, once again the same old no-no. (The same would happen with 'version5-3' and 'version5-a', which would become ('version', 5, -3) and ('version', 5, '-a')).

As you might expect, the solution to the first issue is to wrap the re.split call in a try: except: block and handle the number specially if a TypeError is raised. The second and third cases could be handled in a “special case” manner, meaning only respond and do something different if these problems are detected. But a less error-prone method is to ensure that the data is correct-by-construction, and this can be done by ensuring that the returned tuples always start with a string, and then alternate in a string-number-string-number-string pattern; this can be achieved by adding an empty string wherever the pattern is not followed [2]. This ends up working out pretty nicely because empty strings are always “less” than any non-empty string, and we typically want numbers to come before strings.

Let’s take a look at how this works out.

>>> from natsort.utils import sep_inserter
>>> list(sep_inserter(iter(['apples']), ''))
['apples']
>>>
>>> list(sep_inserter(iter([12, ' apples']), ''))
['', 12, ' apples']
>>>
>>> list(sep_inserter(iter(['version', 5, -3]), ''))
['version', 5, '', -3]
>>>
>>> from natsort import natsort_keygen, ns
>>> natsort_key_with_good_real_number_support = natsort_keygen(alg=ns.REAL)
>>>
>>> sorted([5, '4'], key=natsort_key_with_good_real_number_support)
['4', 5]
>>>
>>> sorted(['12 apples', 'apples'], key=natsort_key_with_good_real_number_support)
['12 apples', 'apples']
>>>
>>> sorted(['version5.3.0', 'version5.3rc1'], key=natsort_key_with_good_real_number_support)
['version5.3.0', 'version5.3rc1']

How the “good” version works will be given in TL;DR 2 - Handling Crappy, Real-World Input.

Handling NaN

A rather unexpected special case I encountered was sorting collections containing NaN. Let’s see what happens when you try to sort a plain old list of numbers when there is a NaN floating around in there.

>>> danger = [7, float('nan'), 22.7, 19, -14, 59.123, 4]
>>> sorted(danger)
[7, nan, -14, 4, 19, 22.7, 59.123]

Clearly that isn’t correct, and for once it isn’t my fault! It’s hard to compare floating point numbers. By definition, NaN is unorderable to any other number, and is never equal to any other number, including itself.

>>> nan = float('nan')
>>> 5 > nan
False
>>> 5 < nan
False
>>> 5 == nan
False
>>> 5 != nan
True
>>> nan == nan
False
>>> nan != nan
True

The implication of all this for us is that if there is an NaN in the data-set we are trying to sort, the data-set will end up being sorted in two separate yet individually sorted sequences - the one before the NaN, and the one after. This is because the < operation that is used to sort always returns False with NaN.

Because natsort aims to sort sequences in a way that does not surprise the user, keeping this behavior is not acceptable (I don’t require my users to know how NaN will behave in a sorting algorithm). The simplest way to satisfy the “least astonishment” principle is to substitute NaN with some other value. But what value is least astonishing? I chose to replace NaN with \(-\infty\) so that these poorly behaved elements always end up at the front where the users will most likely be alerted to their presence.

>>> def fix_nan(x):
...     if x != x:  # only true for NaN
...         return float('-inf')
...     else:
...         return x
...

Let’s check out TL;DR 2 to see how this can be incorporated into the simple key function from TL;DR 1.

TL;DR 2 - Handling Crappy, Real-World Input

Let’s see how our elegant key function from TL;DR 1 has become bastardized in order to support handling mixed real-world data and user customizations.

>>> def natsort_key(x, as_float=False, signed=False, as_path=False):
...     if as_float:
...         regex = signed_float if signed else unsigned_float
...     else:
...         regex = signed_int if signed else unsigned_int
...     try:
...         if as_path:
...             x = decompose_path_into_components(x)  # Decomposes into list of strings
...         # If this raises a TypeError, input is not a string.
...         split_input = re.split(regex, x)
...     except TypeError:
...         try:
...             # Does this need to be applied recursively (list-of-list)?
...             return tuple(map(natsort_key, x))
...         except TypeError:
...             # Must be a number
...             ret = ('', fix_nan(x))  # Maintain string-number-string pattern
...             return (ret,) if as_path else ret  # as_path returns tuple-of-tuples
...     else:
...         split_input = filter(None, split_input)  # removes null strings
...         # Note that the coerce_to_int/coerce_to_float functions
...         # are also modified to use the fix_nan function.
...         if as_float:
...             coerced_input = (coerce_to_float(s) for s in split_input)
...         else:
...             coerced_input = (coerce_to_int(s) for s in split_input)
...         return tuple(sep_inserter(coerced_input, ''))
...

And this doesn’t even show handling bytes type! Notice that we have to do non-obvious things like modify the return form of numbers when as_path is given, just to avoid comparing strings and numbers for the case in which a user provides input like ['/home/me', 42].

Let’s take it out for a spin!

>>> danger = [7, float('nan'), 22.7, '19', '-14', '59.123', 4]
>>> sorted(danger, key=lambda x: natsort_key(x, as_float=True, signed=True))
[nan, '-14', 4, 7, '19', 22.7, '59.123']
>>>
>>> paths = ['Folder (1)/file.tar.gz',
...          'Folder/file.tar.gz',
...          123456]
>>> sorted(paths, key=lambda x: natsort_key(x, as_path=True))
[123456, 'Folder/file.tar.gz', 'Folder (1)/file.tar.gz']

Here Be Dragons: Adding Locale Support

Probably the most challenging special case I had to handle was getting natsort to handle sorting the non-numerical parts of input correctly, and also allowing it to sort the numerical bits in different locales. This was in no way what I originally set out to do with this library, so I was caught a bit off guard when the request was initially made. I discovered the locale library, and assumed that if it’s part of Python’s StdLib there can’t be too many dragons, right?

INCOMPLETE LIST OF DRAGONS

• https://github.com/SethMMorton/natsort/issues/21
• https://github.com/SethMMorton/natsort/issues/22
• https://github.com/SethMMorton/natsort/issues/23
• https://github.com/SethMMorton/natsort/issues/36
• https://github.com/SethMMorton/natsort/issues/44
• https://bugs.python.org/issue2481
• https://bugs.python.org/issue23195
• https://stackoverflow.com/questions/3412933/python-not-sorting-unicode-properly-strcoll-doesnt-help
• https://stackoverflow.com/questions/22203550/sort-dictionary-by-key-using-locale-collation
• https://stackoverflow.com/questions/33459384/unicode-character-not-in-range-when-calling-locale-strxfrm
• https://stackoverflow.com/questions/36431810/sort-numeric-lines-with-thousand-separators
• https://stackoverflow.com/questions/45734562/how-can-i-get-a-reasonable-string-sorting-with-python

These can be summed up as follows:

1. locale is a thin wrapper over your operating system’s locale library, so if that is broken (like it is on BSD and OSX) then locale is broken in Python.
2. Because of a bug in legacy Python (i.e. Python 2), there is no uniform way to use the locale sorting functionality between legacy Python and Python 3.
3. People have differing opinions of how capitalization should affect word order.
4. There is no built-in way to handle locale-dependent thousands separators and decimal points robustly.
5. Proper handling of Unicode is complicated.
6. Proper handling of locale is complicated.

Easily over half of the code in natsort is in some way dealing with some aspect of locale or basic case handling. It would have been impossible to get right without a really good testing strategy.

Don’t expect any more TL;DR’s… if you want to see how all this is fully incorporated into the natsort algorithm then please take a look at the code. However, I will hint at how specific steps are taken in each section.

Let’s see how we can handle some of the dragons, one-by-one.

Basic Case Control Support

Without even thinking about the mess that is adding locale support, natsort can introduce support for controlling how case is interpreted.

First, let’s take a look at how it is sorted by default (due to where characters lie on the ASCII table).

>>> a = ['Apple', 'corn', 'Corn', 'Banana', 'apple', 'banana']
>>> sorted(a)
['Apple', 'Banana', 'Corn', 'apple', 'banana', 'corn']

All uppercase letters come before lowercase letters in the ASCII table, so all capitalized words appear first. Not everyone agrees that this is the correct order. Some believe that the capitalized words should be last (['apple', 'banana', 'corn', 'Apple', 'Banana', 'Corn']). Some believe that both the lowercase and uppercase versions should appear together (['Apple', 'apple', 'Banana', 'banana', 'Corn', 'corn']). Some believe that both should be true ☹. Some people don’t care at all [3].

Solving the first case (I call it LOWERCASEFIRST) is actually pretty easy… just call the str.swapcase() method on the input.

>>> sorted(a, key=lambda x: x.swapcase())
['apple', 'banana', 'corn', 'Apple', 'Banana', 'Corn']

The last (i call it IGNORECASE) should be super easy, right? Simply call str.lowercase() on the input. This will work but may not always give the correct answer on non-latin character sets. It’s a good thing that in Python 3.3 str.casefold() was introduced, which does a better job of removing all case information from unicode characters in non-latin alphabets.

>>> def remove_case(x):
...     try:
...         return x.casefold()
...     except AttributeError:  # Legacy Python backwards compatibility
...         return x.lowercase()
...
>>> sorted(a, key=remove_case)
['Apple', 'apple', 'Banana', 'banana', 'corn', 'Corn']

The middle case (I call it GROUPLETTERS) is less straightforward. The most efficient way to handle this is to duplicate each character with its lowercase version and then the original character.

>>> import itertools
>>> def groupletters(x):
...     return ''.join(itertools.chain.from_iterable((remove_case(y), y) for y in x))
...
>>> groupletters('Apple')
'aAppppllee'
>>> groupletters('apple')
'aappppllee'
>>> sorted(a, key=groupletters)
['Apple', 'apple', 'Banana', 'banana', 'Corn', 'corn']

The effect of this is that both 'Apple' and 'apple' are placed adjacent to each other because their transformations both begin with 'a', and then the second character can be used to order them appropriately with respect to each other.

There’s a problem with this, though. Within the context of natsort we are trying to correctly sort numbers and those should be left alone.

>>> a = ['Apple5', 'apple', 'Apple4E10', 'Banana']
>>> sorted(a, key=lambda x: natsort_key(x, as_float=True))
['Apple5', 'Apple4E10', 'Banana', 'apple']
>>> sorted(a, key=lambda x: natsort_key(groupletters(x), as_float=True))
['Apple4E10', 'Apple5', 'apple', 'Banana']
>>> groupletters('Apple4E10')
'aAppppllee44eE1100'

We messed up the numbers! Looks like groupletters() needs to be applied after the strings are broken into their components. I’m not going to show how this is done here, but basically it requires applying the function in the else: block of coerce_to_int()/coerce_to_float().

>>> better_groupletters = natsort_keygen(alg=ns.GROUPLETTERS | ns.REAL)
>>> better_groupletters('Apple4E10')
('aAppppllee', 40000000000.0)
>>> sorted(a, key=better_groupletters)
['Apple5', 'Apple4E10', 'apple', 'Banana']

Of course, applying both LOWERCASEFIRST and GROUPLETTERS is just a matter of turning on both functions.

Basic Unicode Support

Unicode is hard and complicated. Here’s an example.

>>> b = [b'\x66', b'\x65', b'\xc3\xa9', b'\x65\xcc\x81', b'\x61', b'\x7a']
>>> a = [x.decode('utf8') for x in b]
>>> a  
['f', 'e', 'é', 'é', 'a', 'z']
>>> sorted(a)  
['a', 'e', 'é', 'f', 'z', 'é']

There are more than one way to represent the character ‘é’ in Unicode. In fact, many characters have multiple representations. This is a challenge because comparing the two representations would return False even though they look the same.

>>> a[2] == a[3]
False

Alas, since characters are compared based on the numerical value of their representation, sorting Unicode often gives unexpected results (like seeing ‘é’ come both before and after ‘z’).

The original approach that natsort took with respect to non-ASCII Unicode characters was to say “just use the locale or PyICU library” and then cross it’s fingers and hope those libraries take care of it. As you will find in the following sections, that comes with its own baggage, and turned out to not always work anyway (see https://stackoverflow.com/q/45734562/1399279). A more robust approach is to handle the Unicode out-of-the-box without invoking a heavy-handed library like locale or PyICU. To do this, we must use normalization.

To fully understand Unicode normalization, check out some official Unicode documentation. Just kidding… that’s too much text. The following StackOverflow answers do a good job at explaining Unicode normalization in simple terms: https://stackoverflow.com/a/7934397/1399279 and https://stackoverflow.com/a/7931547/1399279. Put simply, normalization ensures that Unicode characters with multiple representations are in some canonical and consistent representation so that (for example) comparisons of the characters can be performed in a sane way. The following discussion assumes you at least read the StackOverflow answers.

Looking back at our ‘é’ example, we can see that the two versions were constructed with the byte strings b'\xc3\xa9' and b'\x65\xcc\x81'. The former representation is actually LATIN SMALL LETTER E WITH ACUTE and is a single character in the Unicode standard. This is known as the compressed form and corresponds to the ‘NFC’ normalization scheme. The latter representation is actually the letter ‘e’ followed by COMBINING ACUTE ACCENT and so is two characters in the Unicode standard. This is known as the decompressed form and corresponds to the ‘NFD’ normalization scheme. Since the first character in the decompressed form is actually the letter ‘e’, when compared to other ASCII characters it fits where you might expect. Unfortunately, all Unicode compressed form characters come after the ASCII characters and so they always will be placed after ‘z’ when sorting.

It seems that most Unicode data is stored and shared in the compressed form which makes it challenging to sort. This can be solved by normalizing all incoming Unicode data to the decompressed form (‘NFD’) and then sorting.

>>> import unicodedata
>>> c = [unicodedata.normalize('NFD', x) for x in a]
>>> c  
['f', 'e', 'é', 'é', 'a', 'z']
>>> sorted(c)  
['a', 'e', 'é', 'é', 'f', 'z']

Huzzah! Sane sorting without having to resort to locale!

Using Locale to Compare Strings

The locale module is actually pretty cool, and provides lowly spare-time programmers like myself a way to handle the daunting task of proper locale-dependent support of their libraries and utilities. Having said that, it can be a bit of a bear to get right, although they do point out in the documentation that it will be painful to use. Aside from the caveats spelled out in that link, it turns out that just comparing strings with locale in a cross-platform and cross-python-version manner is not as straightforward as one might hope.

First, how to use locale to compare strings? It’s actually pretty straightforward. Simply run the input through the locale transformation function locale.strxfrm().

>>> import locale, sys
>>> locale.setlocale(locale.LC_ALL, 'en_US.UTF-8')
'en_US.UTF-8'
>>> a = ['a', 'b', 'ä']
>>> sorted(a)
['a', 'b', 'ä']
>>> # The below fails on OSX, so don't run doctest on darwin.
>>> is_osx = sys.platform == 'darwin'
>>> sorted(a, key=locale.strxfrm) if not is_osx else ['a', 'ä', 'b']
['a', 'ä', 'b']
>>>
>>> a = ['apple', 'Banana', 'banana', 'Apple']
>>> sorted(a, key=locale.strxfrm) if not is_osx else ['apple', 'Apple', 'banana', 'Banana']
['apple', 'Apple', 'banana', 'Banana']

It turns out that locale-aware sorting groups numbers in the same way as turning on GROUPLETTERS and LOWERCASEFIRST. The trick is that you have to apply locale.strxfrm() only to non-numeric characters; otherwise, numbers won’t be parsed properly. Therefore, it must be applied as part of the coerce_to_int()/coerce_to_float() functions in a manner similar to groupletters().

As you might have guessed, there is a small problem. It turns out the there is a bug in the legacy Python implementation of locale.strxfrm() that causes it to outright fail for unicode() input (https://bugs.python.org/issue2481). locale.strcoll() works, but is intended for use with cmp, which does not exist in current Python implementations. Luckily, the functools.cmp_to_key() function makes locale.strcoll() behave like locale.strxfrm().

Handling Broken Locale On OSX

But what if the underlying locale implementation that locale relies upon is simply broken? It turns out that the locale library on OSX (and other BSD systems) is broken (and for some reason has never been fixed?), and so locale does not work as expected.

How do I define doesn’t work as expected?

>>> a = ['apple', 'Banana', 'banana', 'Apple']
>>> sorted(a)
['Apple', 'Banana', 'apple', 'banana']
>>>
>>> sorted(a, key=locale.strxfrm) if is_osx else sorted(a)
['Apple', 'Banana', 'apple', 'banana']

IT’S SORTING AS IF locale.stfxfrm() WAS NEVER USED!! (and it’s worse once non-ASCII characters get thrown into the mix.) I’m really not sure why this is considered OK for the OSX/BSD maintainers to not fix, but it’s more than frustrating for poor developers who have been dragged into the locale game kicking and screaming. <deep breath>.

So, how to deal with this situation? There are two ways to do so.

1. Detect if locale is sorting incorrectly (i.e. dumb) by seeing if 'A' is sorted before 'a' (incorrect) or not.

   >>> # This is genuinely the name of this function.
   >>> # See natsort.compat.locale.py
   >>> def dumb_sort():
   ...     return locale.strxfrm('A') < locale.strxfrm('a')
   ...
   

   If a dumb locale implementation is found, then automatically turn on LOWERCASEFIRST and GROUPLETTERS.

2. Use an alternate library if installed. ICU is a great and powerful library that has a pretty decent Python port called (you guessed it) PyICU. If a user has this library installed on their computer, natsort chooses to use that instead of locale. With a little bit of planning, one can write a set of wrapper functions that call the correct library under the hood such that the business logic never has to know what library is being used (see natsort.compat.locale.py).

Let me tell you, this little complication really makes a challenge of testing the code, since one must set up different environments on different operating systems in order to test all possible code paths. Not to mention that certain checks will fail for certain operating systems and environments so one must be diligent in either writing the tests not to fail, or ignoring those tests when on offending environments.

Handling Locale-Aware Numbers

Thousands separator support is a problem that I knew would someday be requested but had decided to push off until a rainy day. One day it finally rained, and I decided to tackle the problem.

So what is the problem? Consider the number 1,234,567 (assuming the ',' is the thousands separator). Try to run that through int() and you will get a ValueError. To handle this properly the thousands separators must be removed.

>>> float('1,234,567'.replace(',', ''))
1234567.0

What if, in our current locale, the thousands separator is '.' and the ',' is the decimal separator (like for the German locale de_DE)?

>>> float('1.234.567'.replace('.', '').replace(',', '.'))
1234567.0
>>> float('1.234.567,89'.replace('.', '').replace(',', '.'))
1234567.89

This is pretty much what locale.atoi() and locale.atof() do under the hood. So what’s the problem? Why doesn’t natsort just use this method under its hood? Well, let’s take a look at what would happen if we send some possible natsort input through our the above function:

>>> natsort_key('1,234 apples, please.'.replace(',', ''))
('', 1234, ' apples please.')
>>> natsort_key('Sir, €1.234,50 please.'.replace('.', '').replace(',', '.'), as_float=True)
('Sir. €', 1234.5, ' please')

Any character matching the thousands separator was dropped, and anything matching the decimal separator was changed to '.'! If these characters were critical to how your data was ordered, this would break natsort.

The first solution one might consider would be to first decompose the input into sub-components (like we did for the GROUPLETTERS method above) and then only apply these transformations on the number components. This is a chicken-and-egg problem, though, because we cannot appropriately separate out the numbers because of the thousands separators and non-‘.’ decimal separators (well, at least not without making multiple passes over the data which I do not consider to be a valid option).

Regular expressions to the rescue! With regular expressions, we can remove the thousands separators and change the decimal separator only when they are actually within a number. Once the input has been pre-processed with this regular expression, all the infrastructure shown previously will work.

Beware, these regular expressions will make your eyes bleed.

>>> decimal = ','  # Assume German locale, so decimal separator is ','
>>> # Look-behind assertions cannot accept range modifiers, so instead of i.e.
>>> # (?<!\.[0-9]{1,3}) I have to repeat the look-behind for 1, 2, and 3.
>>> nodecimal = r'(?<!{dec}[0-9])(?<!{dec}[0-9]{{2}})(?<!{dec}[0-9]{{3}})'.format(dec=decimal)
>>> strip_thousands = r'''
...     (?<=[0-9]{{1}})  # At least 1 number
...     (?<![0-9]{{4}})  # No more than 3 numbers
...     {nodecimal}      # Cannot follow decimal
...     {thou}           # The thousands separator
...     (?=[0-9]{{3}}    # Three numbers must follow
...      ([^0-9]|$)      # But a non-number after that
...     )
... '''.format(nodecimal=nodecimal, thou=re.escape('.'))  # Thousands separator is '.' in German locale.
...
>>> re.sub(strip_thousands, '', 'Sir, €1.234,50 please.', flags=re.X)
'Sir, €1234,50 please.'
>>>
>>> # The decimal point must be preceded by a number or after
>>> # a number. This option only needs to be performed in the
>>> # case when the decimal separator for the locale is not '.'.
>>> switch_decimal = r'(?<=[0-9]){decimal}|{decimal}(?=[0-9])'
>>> switch_decimal = switch_decimal.format(decimal=decimal)
>>> re.sub(switch_decimal, '.', 'Sir, €1234,50 please.', flags=re.X)
'Sir, €1234.50 please.'
>>>
>>> natsort_key('Sir, €1234.50 please.', as_float=True)
('Sir, €', 1234.5, ' please.')

Final Thoughts

My hope is that users of natsort never have to think about or worry about all the bookkeeping or any of the details described above, and that using natsort seems to magically “just work”. For those of you who took the time to read this engineering description, I hope it has enlightened you to some of the issues that can be encountered when code is released into the wild and has to accept “real-world data”, or to what happens to developers who naïvely make bold assumptions that are counter to what the rest of the world assumes.

Footnotes

[1]

“But if you hadn’t removed the leading empty string from re.split this wouldn’t have happened!!” I can hear you saying. Well, that’s true. I don’t have a great reason for having done that except that in an earlier non-optimal incarnation of the algorithm I needed to it, and it kind of stuck, and it made other parts of the code easier if the assumption that there were no empty strings was valid.

[2]	I’m not going to show how this is implemented in this document, but if you are interested you can look at the code to sep_inserter() in util.py.

[3]

Handling each of these is straightforward, but coupled with the rapidly fracturing execution paths presented in TL;DR 2 one can imagine this will get out of hand quickly. If you take a look at natsort.py and util.py you can observe that to avoid this I take a more functional approach to construting the natsort algorithm as opposed to the procedural approach illustrated in TL;DR 1 and TL;DR 2.

Examples and Recipes

If you want more detailed examples than given on this page, please see https://github.com/SethMMorton/natsort/tree/master/tests.

Basic Usage

In the most basic use case, simply import natsorted() and use it as you would sorted():

>>> a = ['2 ft 7 in', '1 ft 5 in', '10 ft 2 in', '2 ft 11 in', '7 ft 6 in']
>>> sorted(a)
['1 ft 5 in', '10 ft 2 in', '2 ft 11 in', '2 ft 7 in', '7 ft 6 in']
>>> from natsort import natsorted, ns
>>> natsorted(a)
['1 ft 5 in', '2 ft 7 in', '2 ft 11 in', '7 ft 6 in', '10 ft 2 in']

Sort Version Numbers

As of natsort version >= 4.0.0, natsorted() will work for well-behaved version numbers, like MAJOR.MINOR.PATCH.

Sorting More Expressive Versioning Schemes

By default, if you wish to sort versions that are not as simple as MAJOR.MINOR.PATCH (or similar), you may not get the results you expect:

>>> a = ['1.2', '1.2rc1', '1.2beta2', '1.2beta1', '1.2alpha', '1.2.1', '1.1', '1.3']
>>> natsorted(a)
['1.1', '1.2', '1.2.1', '1.2alpha', '1.2beta1', '1.2beta2', '1.2rc1', '1.3']

To make the ‘1.2’ pre-releases come before ‘1.2.1’, you need to use the following recipe:

>>> natsorted(a, key=lambda x: x.replace('.', '~'))
['1.1', '1.2', '1.2alpha', '1.2beta1', '1.2beta2', '1.2rc1', '1.2.1', '1.3']

If you also want ‘1.2’ after all the alpha, beta, and rc candidates, you can modify the above recipe:

>>> natsorted(a, key=lambda x: x.replace('.', '~')+'z')
['1.1', '1.2alpha', '1.2beta1', '1.2beta2', '1.2rc1', '1.2', '1.2.1', '1.3']

Please see this issue to see why this works.

Sorting Rigorously Defined Versioning Schemes (e.g. SemVer or PEP 440)

If you know you are using a versioning scheme that follows a well-defined format for which there is third-party module support, you should use those modules to assist in sorting. Some examples might be PEP 440 or SemVer.

If we are being honest, using these methods to parse a version means you don’t need to use natsort - you should probably just use sorted() directly. Here’s an example with SemVer:

>>> from semver import VersionInfo
>>> a = ['3.4.5-pre.1', '3.4.5', '3.4.5-pre.2+build.4']
>>> sorted(a, key=VersionInfo.parse)
['3.4.5-pre.1', '3.4.5-pre.2+build.4', '3.4.5']

Sort OS-Generated Paths

In some cases when sorting file paths with OS-Generated names, the default natsorted algorithm may not be sufficient. In cases like these, you may need to use the ns.PATH option:

>>> a = ['./folder/file (1).txt',
...      './folder/file.txt',
...      './folder (1)/file.txt',
...      './folder (10)/file.txt']
>>> natsorted(a)
['./folder (1)/file.txt', './folder (10)/file.txt', './folder/file (1).txt', './folder/file.txt']
>>> natsorted(a, alg=ns.PATH)
['./folder/file.txt', './folder/file (1).txt', './folder (1)/file.txt', './folder (10)/file.txt']

Locale-Aware Sorting (Human Sorting)

Note

Please read Possible Issues with humansorted() or ns.LOCALE before using ns.LOCALE, humansorted(), or index_humansorted().

You can instruct natsort to use locale-aware sorting with the ns.LOCALE option. In addition to making this understand non-ASCII characters, it will also properly interpret non-‘.’ decimal separators and also properly order case. It may be more convenient to just use the humansorted() function:

>>> from natsort import humansorted
>>> import locale
>>> locale.setlocale(locale.LC_ALL, 'en_US.UTF-8')
'en_US.UTF-8'
>>> a = ['Apple', 'corn', 'Corn', 'Banana', 'apple', 'banana']
>>> natsorted(a, alg=ns.LOCALE)
['apple', 'Apple', 'banana', 'Banana', 'corn', 'Corn']
>>> humansorted(a)
['apple', 'Apple', 'banana', 'Banana', 'corn', 'Corn']

You may find that if you do not explicitly set the locale your results may not be as you expect… I have found that it depends on the system you are on. If you use PyICU (see below) then you should not need to do this.

Controlling Case When Sorting

For non-numbers, by default natsort used ordinal sorting (i.e. it sorts by the character’s value in the ASCII table). For example:

>>> a = ['Apple', 'corn', 'Corn', 'Banana', 'apple', 'banana']
>>> natsorted(a)
['Apple', 'Banana', 'Corn', 'apple', 'banana', 'corn']

There are times when you wish to ignore the case when sorting, you can easily do this with the ns.IGNORECASE option:

>>> natsorted(a, alg=ns.IGNORECASE)
['Apple', 'apple', 'Banana', 'banana', 'corn', 'Corn']

Note thats since Python’s sorting is stable, the order of equivalent elements after lowering the case is the same order they appear in the original list.

Upper-case letters appear first in the ASCII table, but many natural sorting methods place lower-case first. To do this, use ns.LOWERCASEFIRST:

>>> natsorted(a, alg=ns.LOWERCASEFIRST)
['apple', 'banana', 'corn', 'Apple', 'Banana', 'Corn']

It may be undesirable to have the upper-case letters grouped together and the lower-case letters grouped together; most would expect all “a”s to bet together regardless of case, and all “b”s, and so on. To achieve this, use ns.GROUPLETTERS:

>>> natsorted(a, alg=ns.GROUPLETTERS)
['Apple', 'apple', 'Banana', 'banana', 'Corn', 'corn']

You might combine this with ns.LOWERCASEFIRST to get what most would expect to be “natural” sorting:

>>> natsorted(a, alg=ns.G | ns.LF)
['apple', 'Apple', 'banana', 'Banana', 'corn', 'Corn']

Customizing Float Definition

You can make natsorted() search for any float that would be a valid Python float literal, such as 5, 0.4, -4.78, +4.2E-34, etc. using the ns.FLOAT key. You can disable the exponential component of the number with ns.NOEXP.

>>> a = ['a50', 'a51.', 'a+50.4', 'a5.034e1', 'a+50.300']
>>> natsorted(a, alg=ns.FLOAT)
['a50', 'a5.034e1', 'a51.', 'a+50.300', 'a+50.4']
>>> natsorted(a, alg=ns.FLOAT | ns.SIGNED)
['a50', 'a+50.300', 'a5.034e1', 'a+50.4', 'a51.']
>>> natsorted(a, alg=ns.FLOAT | ns.SIGNED | ns.NOEXP)
['a5.034e1', 'a50', 'a+50.300', 'a+50.4', 'a51.']

For convenience, the ns.REAL option is provided which is a shortcut for ns.FLOAT | ns.SIGNED and can be used to sort on real numbers. This can be easily accessed with the realsorted() convenience function. Please note that the behavior of the realsorted() function was the default behavior of natsorted() for natsort version < 4.0.0:

>>> natsorted(a, alg=ns.REAL)
['a50', 'a+50.300', 'a5.034e1', 'a+50.4', 'a51.']
>>> from natsort import realsorted
>>> realsorted(a)
['a50', 'a+50.300', 'a5.034e1', 'a+50.4', 'a51.']

Using a Custom Sorting Key

Like the built-in sorted function, natsorted can accept a custom sort key so that:

>>> from operator import attrgetter, itemgetter
>>> a = [['a', 'num4'], ['b', 'num8'], ['c', 'num2']]
>>> natsorted(a, key=itemgetter(1))
[['c', 'num2'], ['a', 'num4'], ['b', 'num8']]
>>> class Foo:
...    def __init__(self, bar):
...        self.bar = bar
...    def __repr__(self):
...        return "Foo('{}')".format(self.bar)
>>> b = [Foo('num3'), Foo('num5'), Foo('num2')]
>>> natsorted(b, key=attrgetter('bar'))
[Foo('num2'), Foo('num3'), Foo('num5')]

Accounting for Units When Sorting

natsort does not come with any pre-built mechanism to sort units, but you can write your own key to do this. Below, I will demonstrate sorting imperial lengths (e.g. feet an inches), but of course you can extend this to any set of units you need. This example is based on code from this issue, and uses the function natsort.numeric_regex_chooser() to build a regular expression that will parse numbers in the same manner as natsort itself.

>>> import re
>>> import natsort
>>>
>>> # Define how each unit will be transformed
>>> conversion_mapping = {
...         "in": 1,
...         "inch": 1,
...         "inches": 1,
...         "ft": 12,
...         "feet": 12,
...         "foot": 12,
... }
>>>
>>> # This regular expression searches for numbers and units
>>> all_units = "|".join(conversion_mapping.keys())
>>> float_re = natsort.numeric_regex_chooser(natsort.FLOAT | natsort.SIGNED)
>>> unit_finder = re.compile(r"({})\s*({})".format(float_re, all_units), re.IGNORECASE)
>>>
>>> def unit_replacer(matchobj):
...     """
...     Given a regex match object, return a replacement string where units are modified
...     """
...     number = matchobj.group(1)
...     unit = matchobj.group(2)
...     new_number = float(number) * conversion_mapping[unit]
...     return "{} in".format(new_number)
...
>>> # Demo time!
>>> data = ['1 ft', '5 in', '10 ft', '2 in']
>>> [unit_finder.sub(unit_replacer, x) for x in data]
['12.0 in', '5.0 in', '120.0 in', '2.0 in']
>>>
>>> natsort.natsorted(data, key=lambda x: unit_finder.sub(unit_replacer, x))
['2 in', '5 in', '1 ft', '10 ft']

Generating a Natsort Key

If you need to sort a list in-place, you cannot use natsorted(); you need to pass a key to the list.sort() method. The function natsort_keygen() is a convenient way to generate these keys for you:

>>> from natsort import natsort_keygen
>>> a = ['a50', 'a51.', 'a50.4', 'a5.034e1', 'a50.300']
>>> natsort_key = natsort_keygen(alg=ns.FLOAT)
>>> a.sort(key=natsort_key)
>>> a
['a50', 'a50.300', 'a5.034e1', 'a50.4', 'a51.']

natsort_keygen() has the same API as natsorted() (minus the reverse option).

Sorting Multiple Lists According to a Single List

Sometimes you have multiple lists, and you want to sort one of those lists and reorder the other lists according to how the first was sorted. To achieve this you could use the index_natsorted() in combination with the convenience function order_by_index():

>>> from natsort import index_natsorted, order_by_index
>>> a = ['a2', 'a9', 'a1', 'a4', 'a10']
>>> b = [4,    5,    6,    7,    8]
>>> c = ['hi', 'lo', 'ah', 'do', 'up']
>>> index = index_natsorted(a)
>>> order_by_index(a, index)
['a1', 'a2', 'a4', 'a9', 'a10']
>>> order_by_index(b, index)
[6, 4, 7, 5, 8]
>>> order_by_index(c, index)
['ah', 'hi', 'do', 'lo', 'up']

Returning Results in Reverse Order

Just like the sorted() built-in function, you can supply the reverse option to return the results in reverse order:

>>> a = ['a2', 'a9', 'a1', 'a4', 'a10']
>>> natsorted(a, reverse=True)
['a10', 'a9', 'a4', 'a2', 'a1']

Sorting Bytes on Python 3

Python 3 is rather strict about comparing strings and bytes, and this can make it difficult to deal with collections of both. Because of the challenge of guessing which encoding should be used to decode a bytes array to a string, natsort does not try to guess and automatically convert for you; in fact, the official stance of natsort is to not support sorting bytes. Instead, some decoding convenience functions have been provided to you (see Help With Bytes On Python 3) that allow you to provide a codec for decoding bytes through the key argument that will allow natsort to convert byte arrays to strings for sorting; these functions know not to raise an error if the input is not a byte array, so you can use the key on any arbitrary collection of data.

>>> from natsort import as_ascii
>>> a = [b'a', 14.0, 'b']
>>> # On Python 2, natsorted(a) would would work as expected.
>>> # On Python 3, natsorted(a) would raise a TypeError (bytes() < str())
>>> natsorted(a, key=as_ascii) == [14.0, b'a', 'b']
True

Additionally, regular expressions cannot be run on byte arrays, making it so that natsort cannot parse them for numbers. As a result, if you run natsort on a list of bytes, you will get results that are like Python’s default sorting behavior. Of course, you can use the decoding functions to solve this:

>>> from natsort import as_utf8
>>> a = [b'a56', b'a5', b'a6', b'a40']
>>> natsorted(a)  
[b'a40', b'a5', b'a56', b'a6']
>>> natsorted(a, key=as_utf8) == [b'a5', b'a6', b'a40', b'a56']
True

If you need a codec different from ASCII or UTF-8, you can use decoder() to generate a custom key:

>>> from natsort import decoder
>>> a = [b'a56', b'a5', b'a6', b'a40']
>>> natsorted(a, key=decoder('latin1')) == [b'a5', b'a6', b'a40', b'a56']
True

Sorting a Pandas DataFrame

Starting from Pandas version 1.1.0, the sorting methods accept a “key” argument, so you can simply pass natsort_keygen() to the sorting methods and sort:

import pandas as pd
from natsort import natsort_keygen
s = pd.Series(['2 ft 7 in', '1 ft 5 in', '10 ft 2 in', '2 ft 11 in', '7 ft 6 in'])
s.sort_values(key=natsort_keygen())
# 1     1 ft 5 in
# 0     2 ft 7 in
# 3    2 ft 11 in
# 4     7 ft 6 in
# 2    10 ft 2 in
# dtype: object

Similarly, if you need to sort the index there is sort_index of a DataFrame.

If you are on an older version of Pandas, check out please check out this answer on StackOverflow for ways to do this without the key argument to sort_values.

natsort API

Standard API

natsorted()

natsort.natsorted(seq: Iterable[Any], key: Optional[Callable[[Any], Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Any]]]]]] = None, reverse: bool = False, alg: Union[natsort.ns_enum.ns, int] = <ns.DEFAULT: 0>) → List[Any]

Sorts an iterable naturally.

Parameters:

• seq (iterable) – The input to sort.
• key (callable, optional) – A key used to determine how to sort each element of the iterable. It is not applied recursively. It should accept a single argument and return a single value.
• reverse ({{True, False}}, optional) – Return the list in reversed sorted order. The default is False.
• alg (ns enum, optional) – This option is used to control which algorithm natsort uses when sorting. For details into these options, please see the ns class documentation. The default is ns.INT.

Returns:

out – The sorted input.

Return type:

list

See also

natsort_keygen()

Generates the key that makes natural sorting possible.

realsorted()

A wrapper for natsorted(seq, alg=ns.REAL).

humansorted()

A wrapper for natsorted(seq, alg=ns.LOCALE).

index_natsorted()

Returns the sorted indexes from natsorted.

os_sorted()

Sort according to your operating system’s rules.

Examples

Use natsorted just like the builtin sorted:

>>> a = ['num3', 'num5', 'num2']
>>> natsorted(a)
['num2', 'num3', 'num5']

The ns enum

natsort.ns

Enum to control the natsort algorithm.

This class acts like an enum to control the natsort algorithm. The user may select several options simultaneously by or’ing the options together. For example, to choose ns.INT, ns.PATH, and ns.LOCALE, you could do ns.INT | ns.LOCALE | ns.PATH. Each function in the natsort package has an alg option that accepts this enum to allow fine control over how your input is sorted.

Each option has a shortened 1- or 2-letter form.

Note

Please read Possible Issues with humansorted() or ns.LOCALE before using ns.LOCALE.

INT, I (default)

The default - parse numbers as integers.

FLOAT, F

Tell natsort to parse numbers as floats.

UNSIGNED, U (default)

Tell natsort to ignore any sign (i.e. “-” or “+”) to the immediate left of a number. This is the default.

SIGNED, S

Tell natsort to take into account any sign (i.e. “-” or “+”) to the immediate left of a number.

REAL, R

This is a shortcut for ns.FLOAT | ns.SIGNED, which is useful when attempting to sort real numbers.

NOEXP, N

Tell natsort to not search for exponents as part of a float number. For example, with NOEXP the number “5.6E5” would be interpreted as 5.6, “E”, and 5 instead of 560000.

NUMAFTER, NA

Tell natsort to sort numbers after non-numbers. By default numbers will be ordered before non-numbers.

PATH, P

Tell natsort to interpret strings as filesystem paths, so they will be split according to the filesystem separator (i.e. ‘/’ on UNIX, ‘’ on Windows), as well as splitting on the file extension, if any. Without this, lists of file paths like ['Folder/', 'Folder (1)/', 'Folder (10)/'] will not be sorted properly; ‘Folder/’ will be placed at the end, not at the front. It is the same as setting the old as_path option to True.

COMPATIBILITYNORMALIZE, CN

Use the “NFKD” unicode normalization form on input rather than the default “NFD”. This will transform characters such as ‘⑦’ into ‘7’. Please see https://stackoverflow.com/a/7934397/1399279, https://stackoverflow.com/a/7931547/1399279, and https://unicode.org/reports/tr15/ for full details into unicode normalization.

LOCALE, L

Tell natsort to be locale-aware when sorting. This includes both proper sorting of alphabetical characters as well as proper handling of locale-dependent decimal separators and thousands separators. This is a shortcut for ns.LOCALEALPHA | ns.LOCALENUM. Your sorting results will vary depending on your current locale.

LOCALEALPHA, LA

Tell natsort to be locale-aware when sorting, but only for alphabetical characters.

LOCALENUM, LN

Tell natsort to be locale-aware when sorting, but only for decimal separators and thousands separators.

IGNORECASE, IC

Tell natsort to ignore case when sorting. For example, ['Banana', 'apple', 'banana', 'Apple'] would be sorted as ['apple', 'Apple', 'Banana', 'banana'].

LOWERCASEFIRST, LF

Tell natsort to put lowercase letters before uppercase letters when sorting. For example, ['Banana', 'apple', 'banana', 'Apple'] would be sorted as ['apple', 'banana', 'Apple', 'Banana'] (the default order would be ['Apple', 'Banana', 'apple', 'banana'] which is the order from a purely ordinal sort). Useless when used with IGNORECASE. Please note that if used with LOCALE, this actually has the reverse effect and will put uppercase first (this is because LOCALE already puts lowercase first); you may use this to your advantage if you need to modify the order returned with LOCALE.

GROUPLETTERS, G

Tell natsort to group lowercase and uppercase letters together when sorting. For example, ['Banana', 'apple', 'banana', 'Apple'] would be sorted as ['Apple', 'apple', 'Banana', 'banana']. Useless when used with IGNORECASE; use with LOWERCASEFIRST to reverse the order of upper and lower case. Generally not needed with LOCALE.

CAPITALFIRST, C

Only used when LOCALE is enabled. Tell natsort to put all capitalized words before non-capitalized words. This is essentially the inverse of GROUPLETTERS, and is the default Python sorting behavior without LOCALE.

UNGROUPLETTERS, UG

An alias for CAPITALFIRST.

NANLAST, NL

If an NaN shows up in the input, this instructs natsort to treat these as +Infinity and place them after all the other numbers. By default, an NaN be treated as -Infinity and be placed first. Note that this None is treated like NaN internally.

Notes

If you prefer to use import natsort as ns as opposed to from natsort import natsorted, ns, the ns options are available as top-level imports.

>>> import natsort as ns
>>> a = ['num5.10', 'num-3', 'num5.3', 'num2']
>>> ns.natsorted(a, alg=ns.REAL) == ns.natsorted(a, alg=ns.ns.REAL)
True

natsort_key()

natsort.natsort_key(val)

The default natural sorting key.

This is the output of natsort_keygen() with default values.

See also

natsort_keygen()

natsort_keygen()

natsort.natsort_keygen(key: Optional[Callable[[Any], Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Any]]]]]] = None, alg: Union[natsort.ns_enum.ns, int] = <ns.DEFAULT: 0>) → Callable[[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Any]]]]], Tuple[Union[str, bytes, float, int, Tuple[Union[str, bytes, float, int, Tuple[Any, ...]], ...]], ...]]

Generate a key to sort strings and numbers naturally.

This key is designed for use as the key argument to functions such as the sorted builtin.

The user may customize the generated function with the arguments to natsort_keygen, including an optional key function.

Parameters:

• key (callable, optional) – A key used to manipulate the input value before parsing for numbers. It is not applied recursively. It should accept a single argument and return a single value.
• alg (ns enum, optional) – This option is used to control which algorithm natsort uses when sorting. For details into these options, please see the ns class documentation. The default is ns.INT.

Returns:

out – A function that parses input for natural sorting that is suitable for passing as the key argument to functions such as sorted.

Return type:

function

See also

natsorted(), natsort_key()

Examples

natsort_keygen is a convenient way to create a custom key to sort lists in-place (for example).:

>>> a = ['num5.10', 'num-3', 'num5.3', 'num2']
>>> a.sort(key=natsort_keygen(alg=ns.REAL))
>>> a
['num-3', 'num2', 'num5.10', 'num5.3']

os_sort_key()

natsort.os_sort_key(val)

The default key to replicate your file browser’s sort order

This is the output of os_sort_keygen() with default values.

See also

os_sort_keygen()

os_sort_keygen()

natsort.os_sort_keygen(key: Optional[Callable[[Any], Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Any]]]]]] = None) → Callable[[Union[str, bytes, float, int, pathlib.PurePath, None]], Tuple[Union[str, bytes, float, int, Tuple[Union[str, bytes, float, int], ...]], ...]]

Generate a sorting key to replicate your file browser’s sort order

See os_sorted() for description and caveats.

Returns:out – A function that parses input for OS path sorting that is suitable for passing as the key argument to functions such as sorted. Return type:function

See also

os_sort_key(), os_sorted()

Notes

On Windows, this will implicitly coerce all inputs to str before collating.

Convenience Functions

os_sorted()

natsort.os_sorted(seq: Iterable[Any], key: Optional[Callable[[Any], Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Any]]]]]] = None, reverse: bool = False) → List[Any]

Sort elements in the same order as your operating system’s file browser

Warning

The resulting function will generate results that will be differnt depending on your platform. This is intentional.

On Windows, this will sort with the same order as Windows Explorer.

On MacOS/Linux, you will get different results depending on whether or not you have pyicu installed.

• If you have pyicu installed, you will get results that are the same as (or very close to) the same order as your operating system’s file browser.
• If you do not have pyicu installed, then this will give the same results as if you used ns.LOCALE, ns.PATH, and ns.IGNORECASE with natsorted(). If you do not have special characters this will give correct results, but once special characters are added you should lower your expectations.

It is strongly reccommended to have pyicu installed on MacOS/Linux if you want correct sort results.

It does not take into account if a path is a directory or a file when sorting.

Parameters:

• seq (iterable) – The input to sort. Each element must be of type str.
• key (callable, optional) – A key used to determine how to sort each element of the sequence. It should accept a single argument and return a single value.
• reverse ({{True, False}}, optional) – Return the list in reversed sorted order. The default is False.

Returns:

out – The sorted input.

Return type:

list

See also

natsorted(), os_sort_keygen()

Notes

This will implicitly coerce all inputs to str before collating.

realsorted()

natsort.realsorted(seq: Iterable[Any], key: Optional[Callable[[Any], Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Any]]]]]] = None, reverse: bool = False, alg: Union[natsort.ns_enum.ns, int] = <ns.DEFAULT: 0>) → List[Any]

Convenience function to properly sort signed floats.

A signed float in a string could be “a-5.7”. This is a wrapper around natsorted(seq, alg=ns.REAL).

The behavior of realsorted() for natsort version >= 4.0.0 was the default behavior of natsorted() for natsort version < 4.0.0.

Parameters:

• seq (iterable) – The input to sort.
• key (callable, optional) – A key used to determine how to sort each element of the sequence. It is not applied recursively. It should accept a single argument and return a single value.
• reverse ({{True, False}}, optional) – Return the list in reversed sorted order. The default is False.
• alg (ns enum, optional) – This option is used to control which algorithm natsort uses when sorting. For details into these options, please see the ns class documentation. The default is ns.REAL.

Returns:

out – The sorted input.

Return type:

list

See also

index_realsorted()

Returns the sorted indexes from realsorted.

Examples

Use realsorted just like the builtin sorted:

>>> a = ['num5.10', 'num-3', 'num5.3', 'num2']
>>> natsorted(a)
['num2', 'num5.3', 'num5.10', 'num-3']
>>> realsorted(a)
['num-3', 'num2', 'num5.10', 'num5.3']

humansorted()

natsort.humansorted(seq: Iterable[Any], key: Optional[Callable[[Any], Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Any]]]]]] = None, reverse: bool = False, alg: Union[natsort.ns_enum.ns, int] = <ns.DEFAULT: 0>) → List[Any]

Convenience function to properly sort non-numeric characters.

This is a wrapper around natsorted(seq, alg=ns.LOCALE).

Parameters:

• seq (iterable) – The input to sort.
• key (callable, optional) – A key used to determine how to sort each element of the sequence. It is not applied recursively. It should accept a single argument and return a single value.
• reverse ({{True, False}}, optional) – Return the list in reversed sorted order. The default is False.
• alg (ns enum, optional) – This option is used to control which algorithm natsort uses when sorting. For details into these options, please see the ns class documentation. The default is ns.LOCALE.

Returns:

out – The sorted input.

Return type:

list

See also

index_humansorted()

Returns the sorted indexes from humansorted.

Notes

Please read Possible Issues with humansorted() or ns.LOCALE before using humansorted.

Examples

Use humansorted just like the builtin sorted:

>>> a = ['Apple', 'Banana', 'apple', 'banana']
>>> natsorted(a)
['Apple', 'Banana', 'apple', 'banana']
>>> humansorted(a)
['apple', 'Apple', 'banana', 'Banana']

index_natsorted()

natsort.index_natsorted(seq: Iterable[Any], key: Optional[Callable[[Any], Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Any]]]]]] = None, reverse: bool = False, alg: Union[natsort.ns_enum.ns, int] = <ns.DEFAULT: 0>) → List[int]

Determine the list of the indexes used to sort the input sequence.

Sorts a sequence naturally, but returns a list of sorted the indexes and not the sorted list itself. This list of indexes can be used to sort multiple lists by the sorted order of the given sequence.

Parameters:

• seq (iterable) – The input to sort.
• key (callable, optional) – A key used to determine how to sort each element of the sequence. It is not applied recursively. It should accept a single argument and return a single value.
• reverse ({{True, False}}, optional) – Return the list in reversed sorted order. The default is False.
• alg (ns enum, optional) – This option is used to control which algorithm natsort uses when sorting. For details into these options, please see the ns class documentation. The default is ns.INT.

Returns:

out – The ordered indexes of the input.

Return type:

tuple

See also

natsorted(), order_by_index()

Examples

Use index_natsorted if you want to sort multiple lists by the sorted order of one list:

>>> a = ['num3', 'num5', 'num2']
>>> b = ['foo', 'bar', 'baz']
>>> index = index_natsorted(a)
>>> index
[2, 0, 1]
>>> # Sort both lists by the sort order of a
>>> order_by_index(a, index)
['num2', 'num3', 'num5']
>>> order_by_index(b, index)
['baz', 'foo', 'bar']

index_realsorted()

natsort.index_realsorted(seq: Iterable[Any], key: Optional[Callable[[Any], Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Any]]]]]] = None, reverse: bool = False, alg: Union[natsort.ns_enum.ns, int] = <ns.DEFAULT: 0>) → List[int]

This is a wrapper around index_natsorted(seq, alg=ns.REAL).

Parameters:

• seq (iterable) – The input to sort.
• key (callable, optional) – A key used to determine how to sort each element of the sequence. It is not applied recursively. It should accept a single argument and return a single value.
• reverse ({{True, False}}, optional) – Return the list in reversed sorted order. The default is False.
• alg (ns enum, optional) – This option is used to control which algorithm natsort uses when sorting. For details into these options, please see the ns class documentation. The default is ns.REAL.

Returns:

out – The ordered indexes of the input.

Return type:

tuple

See also

realsorted(), order_by_index()

Examples

Use index_realsorted just like the builtin sorted:

>>> a = ['num5.10', 'num-3', 'num5.3', 'num2']
>>> index_realsorted(a)
[1, 3, 0, 2]

index_humansorted()

natsort.index_humansorted(seq: Iterable[Any], key: Optional[Callable[[Any], Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Any]]]]]] = None, reverse: bool = False, alg: Union[natsort.ns_enum.ns, int] = <ns.DEFAULT: 0>) → List[int]

This is a wrapper around index_natsorted(seq, alg=ns.LOCALE).

Parameters:

• seq (iterable) – The input to sort.
• key (callable, optional) – A key used to determine how to sort each element of the sequence. It is not applied recursively. It should accept a single argument and return a single value.
• reverse ({{True, False}}, optional) – Return the list in reversed sorted order. The default is False.
• alg (ns enum, optional) – This option is used to control which algorithm natsort uses when sorting. For details into these options, please see the ns class documentation. The default is ns.LOCALE.

Returns:

out – The ordered indexes of the input.

Return type:

tuple

See also

humansorted(), order_by_index()

Notes

Please read Possible Issues with humansorted() or ns.LOCALE before using humansorted.

Examples

Use index_humansorted just like the builtin sorted:

>>> a = ['Apple', 'Banana', 'apple', 'banana']
>>> index_humansorted(a)
[2, 0, 3, 1]

order_by_index()

natsort.order_by_index(seq: Sequence[Any], index: Iterable[int], iter: bool = False) → Iterable[Any]

Order a given sequence by an index sequence.

The output of index_natsorted is a sequence of integers (index) that correspond to how its input sequence would be sorted. The idea is that this index can be used to reorder multiple sequences by the sorted order of the first sequence. This function is a convenient wrapper to apply this ordering to a sequence.

Parameters:

• seq (sequence) – The sequence to order.
• index (iterable) – The iterable that indicates how to order seq. It should be the same length as seq and consist of integers only.
• iter ({{True, False}}, optional) – If True, the ordered sequence is returned as a iterator; otherwise it is returned as a list. The default is False.

Returns:

out – The sequence ordered by index, as a list or as an iterator (depending on the value of iter).

Return type:

{{list, iterator}}

See also

index_natsorted(), index_humansorted(), index_realsorted()

Examples

order_by_index is a convenience function that helps you apply the result of index_natsorted:

>>> a = ['num3', 'num5', 'num2']
>>> b = ['foo', 'bar', 'baz']
>>> index = index_natsorted(a)
>>> index
[2, 0, 1]
>>> # Sort both lists by the sort order of a
>>> order_by_index(a, index)
['num2', 'num3', 'num5']
>>> order_by_index(b, index)
['baz', 'foo', 'bar']

Help With Bytes On Python 3

The official stance of natsort is to not support bytes for sorting; there is just too much that can go wrong when trying to automate conversion between bytes and str. But rather than completely give up on bytes, natsort provides three functions that make it easy to quickly decode bytes to str so that sorting is possible.

natsort.decoder(encoding: str) → Callable[[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Any]]]]], Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Any]]]]]

Return a function that can be used to decode bytes to unicode.

Parameters:encoding (str) – The codec to use for decoding. This must be a valid unicode codec. Returns:A function that takes a single argument and attempts to decode it using the supplied codec. Any UnicodeErrors are raised. If the argument was not of bytes type, it is simply returned as-is. Return type:decode_function

See also

as_ascii(), as_utf8()

Examples

>>> f = decoder('utf8')
>>> f(b'bytes') == 'bytes'
True
>>> f(12345) == 12345
True
>>> # On Python 3, without decoder this would return [b'a10', b'a2']
>>> natsorted([b'a10', b'a2'], key=decoder('utf8')) == [b'a2', b'a10']
True
>>> # On Python 3, without decoder this would raise a TypeError.
>>> natsorted([b'a10', 'a2'], key=decoder('utf8')) == ['a2', b'a10']
True

natsort.as_ascii(s: Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Any]]]]) → Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Any]]]]

Function to decode an input with the ASCII codec, or return as-is.

Parameters:s (object) – Returns:If the input was of type bytes, the return value is a str decoded with the ASCII codec. Otherwise, the return value is identically the input. Return type:output

See also

decoder()

natsort.as_utf8(s: Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Any]]]]) → Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Union[str, bytes, float, int, pathlib.PurePath, None, Iterable[Any]]]]

Function to decode an input with the UTF-8 codec, or return as-is.

Parameters:s (object) – Returns:If the input was of type bytes, the return value is a str decoded with the UTF-8 codec. Otherwise, the return value is identically the input. Return type:output

See also

decoder()

Help With Creating Function Keys

If you need to create a complicated key argument to (for example) natsorted() that is actually multiple functions called one after the other, the following function can help you easily perform this action. It is used internally to natsort, and has been exposed publicly for the convenience of the user.

natsort.chain_functions(functions: Iterable[Callable[[Any], Any]]) → Callable[[Any], Any]

Chain a list of single-argument functions together and return.

The functions are applied in list order, and the output of the previous functions is passed to the next function.

Parameters:functions (list) – A list of single-argument functions to chain together. Returns:func – A single argument function. Return type:callable

Examples

Chain several functions together!

>>> funcs = [lambda x: x * 4, len, lambda x: x + 5]
>>> func = chain_functions(funcs)
>>> func('hey')
17

If you need to be able to search your input for numbers using the same definition as natsort, you can do so using the following function. Given your chosen algorithm (selected using the ns enum), the corresponding regular expression to locate numbers will be returned.

natsort.numeric_regex_chooser(alg: Union[natsort.ns_enum.ns, int]) → str

Select an appropriate regex for the type of number of interest.

Parameters:alg (ns enum) – Used to indicate the regular expression to select. Returns:regex – Regular expression string that matches the desired number type. Return type:str

Help With Type Hinting

If you need to explictly specify the types that natsort accepts or returns in your code, the following types have been exposed for your convenience.

Type	Purpose
natsort.NatsortKeyType	Returned by natsort.natsort_keygen(), and type of natsort.natsort_key
natsort.OSSortKeyType	Returned by natsort.os_sort_keygen(), and type of natsort.os_sort_key
natsort.KeyType	Type of key argument to natsort.natsorted() and natsort.natsort_keygen()
natsort.NatsortInType	The input type of natsort.NatsortKeyType
natsort.NatsortOutType	The output type of natsort.NatsortKeyType
natsort.NSType	The type of the ns enum

Possible Issues with humansorted() or ns.LOCALE

Being Locale-Aware Means Both Numbers and Non-Numbers

In addition to modifying how characters are sorted, ns.LOCALE will take into account locale-dependent thousands separators (and locale-dependent decimal separators if ns.FLOAT is enabled). This means that if you are in a locale that uses commas as the thousands separator, a number like 123,456 will be interpreted as 123456. If this is not what you want, you may consider using ns.LOCALEALPHA which will only enable locale-aware sorting for non-numbers (similarly, ns.LOCALENUM enables locale-aware sorting only for numbers).

Regenerate Key With natsort_keygen() After Changing Locale

When natsort_keygen() is called it returns a key function that hard-codes the provided settings. This means that the key returned when ns.LOCALE is used contains the settings specifed by the locale loaded at the time the key is generated. If you change the locale, you should regenerate the key to account for the new locale.

Corollary: Do Not Reuse natsort_keygen() After Changing Locale

If you change locale, the old function will not work as expected. The locale library works with a global state. When natsort_keygen() is called it does the best job that it can to make the returned function as static as possible and independent of the global state, but the locale.strxfrm() function must access this global state to work; therefore, if you change locale and use ns.LOCALE then you should discard the old key.

Note

If you use PyICU then you may be able to reuse keys after changing locale.

The locale Module From the StdLib Has Issues

natsort will use PyICU for humansorted() or ns.LOCALE if it is installed. If not, it will fall back on the locale library from the Python stdlib. If you do not have PyICU installed, please keep the following known problems and issues in mind.

Note

Remember, if you have PyICU installed you shouldn’t need to worry about any of these.

Explicitly Set the Locale Before Using ns.LOCALE

I have found that unless you explicitly set a locale, the sorted order may not be what you expect. Setting this is straightforward (in the below example I use ‘en_US.UTF-8’, but you should use your locale):

>>> import locale
>>> locale.setlocale(locale.LC_ALL, 'en_US.UTF-8')
'en_US.UTF-8'

The locale Module Is Broken on Mac OS X

It’s not Python’s fault, but the OS… the locale library for BSD-based systems (of which Mac OS X is one) is broken. See the following links:

• https://stackoverflow.com/questions/3412933/python-not-sorting-unicode-properly-strcoll-doesnt-help
• https://bugs.python.org/issue23195
• https://github.com/SethMMorton/natsort/issues/21 (contains instructons on installing)
• https://stackoverflow.com/questions/33459384/unicode-character-not-in-range-when-calling-locale-strxfrm
• https://github.com/SethMMorton/natsort/issues/34

Of course, installing PyICU fixes this, but if you don’t want to or cannot install this there is some hope.

1. As of natsort version 4.0.0, natsort is configured to compensate for a broken locale library. When sorting non-numbers it will handle case as you expect, but it will still not be able to comprehend non-ASCII characters properly. Additionally, it has a built-in lookup table of thousands separators that are incorrect on OS X/BSD (but is possible it is not complete… please file an issue if you see it is not complete)
2. Use “*.ISO8859-1” locale (i.e. ‘en_US.ISO8859-1’) rather than “*.UTF-8” locale. I have found that these have fewer issues than “UTF-8”, but your mileage may vary.

Shell Script

The natsort shell script is automatically installed when you install natsort with pip.

Below is the usage and some usage examples for the natsort shell script.

Usage

Description

natsort was originally written to aid in computational chemistry research so that it would be easy to analyze large sets of output files named after the parameter used:

$ ls *.out
mode1000.35.out mode1243.34.out mode744.43.out mode943.54.out

(Obviously, in reality there would be more files, but you get the idea.) Notice that the shell sorts in lexicographical order. This is the behavior of programs like find as well as ls. The problem is passing these files to an analysis program causes them not to appear in numerical order, which can lead to bad analysis. To remedy this, use natsort:

$ natsort *.out
mode744.43.out
mode943.54.out
mode1000.35.out
mode1243.34.out
$ natsort -t r *.out | xargs your_program

-t r is short for --number-type real. You can also place natsort in the middle of a pipe:

$ find . -name "*.out" | natsort -t r | xargs your_program

To sort version numbers, use the default --number-type:

$ ls *
prog-1.10.zip prog-1.9.zip prog-2.0.zip
$ natsort *
prog-1.9.zip
prog-1.10.zip
prog-2.0.zip

In general, all natsort shell script options mirror the natsorted() API, with notable exception of the --filter, --reverse-filter, and --exclude options. These three options are used as follows:

$ ls *.out
mode1000.35.out mode1243.34.out mode744.43.out mode943.54.out
$ natsort -t r *.out -f 900 1100 # Select only numbers between 900-1100
mode943.54.out
mode1000.35.out
$ natsort -t r *.out -F 900 1100 # Select only numbers NOT between 900-1100
mode744.43.out
mode1243.34.out
$ natsort -t r *.out -e 1000.35 # Exclude 1000.35 from search
mode744.43.out
mode943.54.out
mode1243.34.out

If you are sorting paths with OS-generated filenames, you may require the --paths/-p option:

$ find . ! -path . -type f
./folder/file (1).txt
./folder/file.txt
./folder (1)/file.txt
./folder (10)/file.txt
./folder (2)/file.txt
$ find . ! -path . -type f | natsort
./folder (1)/file.txt
./folder (2)/file.txt
./folder (10)/file.txt
./folder/file (1).txt
./folder/file.txt
$ find . ! -path . -type f | natsort -p
./folder/file.txt
./folder/file (1).txt
./folder (1)/file.txt
./folder (2)/file.txt
./folder (10)/file.txt

Changelog

Unreleased

8.0.0 - 2021-11-03

• Re-release 7.2.0 as 8.0.0 because introduction of type hints can break CI builds (issue #139)

7.2.0 - 2021-11-02 (Yanked)

Added

• Type hints (contributions from @thethiny and @domdfcoding, issues #132, #135, and #138)
• Explicit testing for Python 3.10

Removed

• Support for Python 3.4 and Python 3.5

7.1.1 - 2021-01-24

Changed

• Use GitHub Actions instead of Travis-CI (issue #125)
• No longer pin testing dependencies (issue #126)

Fixed

• Correct a minor typo (@madphysicist, issue #127)

7.1.0 - 2020-11-19

Added

• os_sorted, os_sort_keygen, and os_sort_key to better support sorting like the file browser on the current operating system - this closes the long-standing issue #41
• Support for Python 3.9 (@swt2c, issue #119)

Changed

• MacOS unit tests run on native Python
• Treate None like NaN internally to avoid TypeError (issue #117)
• No longer fail tests every time a new Python version is released (issue #122)

Fixed

• Various typos, missing figures, and out-of-date information in the “How it works”
• Fix typo in CHANGELOG (@graingert, issue #113)
• Updated “How it works” to account for Pandas updates (@kuraga, issue #116)

7.0.1 - 2020-01-27

Fixed

• Bug where that caused incorrect sorting when using locales that have a "." character as the thousands separator.

7.0.0 - 2020-01-08

Added

• Ability to deploy directly from TravisCI (@hugovk, issue #106)
• Release checklist in RELEASING.md (@hugovk, issue #106)

Changed

• Updated auxillary shell scripts to be written in python, and added ability to call these from tox
• Improved Travis-CI experience
• Update testing dependency versions

Removed

• Support for Python 2

6.2.0 - 2019-11-13

Added

• Support for Python 3.8 (@hugovk, issue #104)

Changed

• index_natsorted internally now uses tuples for index-element pairs instead of lists
• Added a TOC to the README
• Python 3.4 is no longer included in testing

Fixed

• Pin testing dependencies to prevent CI breaking due to third-party library changes

Removed

• Introduction page in documentation

6.1.0 - 2019-11-09

Added

• Expose numeric_regex_chooser as a public function for ease in making key functions
• Example in the documentation on how to sort numbers with units
• Automated testing support for macos and Windows (issue #91)

Changed

• Update CHANGELOG format to style from https://keepachangelog.com/ (issue #92)

Fixed

• Removed dependency on sudo in TravisCI configuration (@hugovk, issue #99)
• Documentation typos (@jdufresne, issue #94) (@cpburnz, issue #95)

6.0.0 - 2019-02-04

Changed

• Simply Travis-CI configuration (@jdufresne, issue #88)

Fixed

• Fix README rendering in PyPI (@altendky, issue #89)

Removed

• Drop support for Python 2.6 and 3.3 (@jdufresne, issue #70)
• Remove deprecated APIs (kwargs number_type, signed, exp, as_path, py3_safe; enums ns.TYPESAFE, ns.DIGIT, ns.VERSION; functions versorted, index_versorted) (issue #81)
• Remove pipenv as a dependency for building (issue #86)

5.5.0 - 2018-11-18

Added

• CHANGELOG.rst to the top-level of the repository (issue #85)

Changed

• Documentation, packaging, and CI cleanup (@jdufresne, issues #69, #71-#80)
• Consolidate API documentation into a single page (issue #82)

Deprecated

• Formally deprecated old or misleading APIs (issue #83)

Fixed

• Add back support for very old versions of setuptools (issue #84)

5.4.1 - 2018-09-09

Changed

• Code format and quality checking infrastructure (issue #68)

Fixed

• Error in a newly added test (issues #65, #67)

5.4.0 - 2018-09-06

Changed

• Re-expose natsort_key as “public” and remove the associated DeprecationWarning
• Better developer documentation
• Refactor tests (issue #66)
• Bump allowed ``fastnumbers` <https://github.com/SethMMorton/fastnumbers>`_ version

5.3.3 - 2018-07-07

Added

• Enable Python 3.7 support in Travis-CI (issue #61)

Changed

• Update docs with a FAQ and quick how-it-works (issue #60)

Fixed

• StopIteration error in the testing code

5.3.2 - 2018-05-17

Fixed

• Bug that prevented install on old versions of setuptools (issues #55, #56)
• Revert layout from src/natsort/ back to natsort/ to make user testing simpler (issues #57, #58)

5.3.1 - 2018-05-14

Added

• ``bumpversion` <https://github.com/c4urself/bump2version>`_ infrastructure
• Extras can be installed by “[]” notation

Changed

• No bugfixes or features, just infrastructure and installation updates
• Move to defining dependencies with Pipfile
• Development layout is now src/natsort/ instead of natsort/

5.3.0 - 2018-04-20

Added

• Ability to consider unicode-decimal numbers as numbers (issues #52, #54)

Fixed

• Bug in assessing ``fastnumbers` <https://github.com/SethMMorton/fastnumbers>`_ version at import-time (@hholzgra, issues #51, #53)

5.2.0 - 2018-02-14

Added

• ns.NUMAFTER to cause numbers to be placed after non-numbers (issues #48, #49)
• natcmp function (Python 2 only) (@rinslow, issue #47)

5.1.1 - 2017-11-11

Added

• Additional unicode number support for Python 3.7
• Information on how to install and test (issue #46)

5.1.0 - 2017-08-19

Changed

• All Unicode input is now normalized (issue #44, #45)

Fixed

• StopIteration warning on Python 3.6+ (@lykinsbd, issues #42, #43)

5.0.3 - 2017-04-30

• Improved development infrastructure
• Migrated documentation to ReadTheDocs

5.0.2 - 2017-01-02

Added

• Additional unicode number support for Python 3.6
• “how does it work?” section to the documentation

Changed

• Renamed several internal functions and variables to improve clarity
• Improved documentation examples

5.0.1 - 2016-06-04

Added

• The ns enum attributes can now be imported from the top-level namespace

Fixed

• Bug with the from natsort import * mechanism
• Bug with using natsort with python -OO (issues #38, #39)

5.0.0 - 2016-05-08

Added

• chain_functions function for convenience in creating a complex user-given key from several existing functions

Changed

• ns.LOCALE/humansorted now accounts for thousands separators (issue #36)
• Refactored entire codebase to be more functional (as in use functions as units). Previously, the code was rather monolithic and difficult to follow. The goal is that with the code existing in smaller units, contributing will be easier (issue #37)
• Increased speed of execution (came for free with the new functional approach because the new factory function paradigm eliminates most if branches during execution). For the most cases, the code is 30-40% faster than version 4.0.4. If using ns.LOCALE or humansorted, the code is 1100% faster than version 4.0.4
• Improved clarity of documentaion with regards to locale-aware sorting

Deprecated

• ns.TYPESAFE option as it is now always on (due to a new iterator-based algorithm, the typesafe function is now cheap)

4.0.4 - 2015-11-01

Changed

• Improved coverage of unit tests
• Unit tests use new and improved hypothesis library

Fixed

• Compatibility issues with Python 3.5

4.0.3 - 2015-06-25

Fixed

• Bad install on last release (sorry guys!) (issue #30)

4.0.2 - 2015-06-24

Changed

• Consolidated under-the-hood compatibility functionality

Fixed

• Python 2.6 and Python 3.2 compatibility. Unit testing is now performed for these versions (@dpetzold, issue #29)

4.0.1 - 2015-06-04

Added

• Support for sorting NaN by internally converting to -Infinity or +Infinity (issue #27)

4.0.0 - 2015-05-17

Changed

• Made default behavior of natsort search for unsigned ints, rather than signed floats. This is a backwards-incompatible change but in 99% of use cases it should not require any end-user changes (issue #20)
• Improved handling of locale-aware sorting on systems where the underlying locale library is broken (issue #34))
• Greatly improved all unit tests by adding the hypothesis library

3.5.6 - 2015-04-06

Added

• UNGROUPLETTERS algorithm to get the case-grouping behavior of an ordinal sort when using LOCALE (issue #23)
• Convenience functions decoder, as_ascii, and as_utf8 for dealing with bytes types

3.5.5 - 2015-04-04

Added

• realsorted and index_realsorted functions for forward-compatibility with >= 4.0.0

Changed

• Made explanation of when to use TYPESAFE more clear in the docs

3.5.4 - 2015-04-02

Fixed

• Bug where a TypeError was raised if a string containing a leading number was sorted with alpha-only strings when LOCALE is used (issue #22)

3.5.3 - 2015-03-26

Changed

• Documentation updates to better describe locale bug, and illustrate upcoming default behavior change
• Internal improvements, including making test suite more granular

Fixed

• Bug where --reverse-filter option in shell script was not getting checked for correctness

3.5.2 - 2015-01-13

Added

• A pathlib.Path object is converted to a str if ns.PATH is enabled (issue #16)

3.5.1 - 2014-09-25

Changed

• Refactored modules so that only the public API was in natsort.py and ns_enum.py
• Refactored all import statements to be absolute, not relative

Fixed

• Bug that caused list/tuples to fail when using ns.LOWECASEFIRST or ns.IGNORECASE (issue #15)

3.5.0 - 2014-09-02

Added

• alg argument to the natsort functions. This argument accepts an enum that is used to indicate the options the user wishes to use. The number_type, signed, exp, as_path, and py3_safe options are being deprecated and will become (undocumented) keyword-only options in natsort version 4.0.0
• The humansorted convenience function as a convenience to locale-aware sorting
• The user can now modify how natsort handles the case of non-numeric characters (issue #14)
• The user can now instruct natsort to use locale-aware sorting, which allows natsort to perform true “human sorting” (issue #14)
• Locale functionality to the shell script

3.4.1 - 2014-08-12

Changed

• natsort will now use the ``fastnumbers` <https://github.com/SethMMorton/fastnumbers>`_ module if it is installed. This gives up to an extra 30% boost in speed over the previous performance enhancements
• Made documentation point to more natsort resources, and also added a new example in the examples section

3.4.0 - 2014-07-19

Added

• natsort_keygen function that will generate a wrapped version of natsort_key that is easier to call. natsort_key is now set to deprecate at natsort version 4.0.0
• as_path option to natsorted & co. that will try to treat input strings as filepaths. This will help yield correct results for OS-generated inputs like ['/p/q/o.x', '/p/q (1)/o.x', '/p/q (10)/o.x', '/p/q/o (1).x'] (issue #3)
• order_by_index function to help in using the output of index_natsorted and index_versorted
• reverse option to natsorted & co. to make it’s API more similar to the builtin ‘sorted’
• More unit tests
• Auxillary test code that helps in profiling and stress-testing
• Support for coveralls.io

Changed

• Massive performance enhancements for string input (1.8x-2.0x), at the expense of reduction in speed for numeric input (~2.0x) - note that sorting numbersstill only takes 0.6x the time of sorting strings
• Entire codebase is now PyFlakes and PEP8 compliant
• Reworked the documentation, moving most of it to PyPI’s hosting platform

Fixed

• Bug that caused user’s options to the natsort_key to not be passed on to recursive calls of natsort_key (issue #12)

3.3.0 - 2014-06-28

Added

• versorted method for more convenient sorting of versions (issue #11)
• Unit test coverage (99%)

Changed

• Updated command-line tool --number_type option with ‘version’ and ‘ver’ to make it more clear how to sort version numbers
• Moved unit-testing mechanism from being docstring-based to actual unit tests in actual functions (issue #10)
• Made docstrings for public functions mirror the README API
• Connected natsort development to Travis-CI to help ensure quality releases

3.2.1 - 2014-06-20

Fixed

• Re-“Fixed” unorderable types issue on Python 3.x - this workaround is for when the problem occurs in the middle of the string (issue #7 again)

3.2.0 - 2014-05-07

Fixed

• “Fixed” unorderable types issue on Python 3.x with a workaround that attempts to replicate the Python 2.x behavior by putting all the numbers (or strings that begin with numbers) first (issue #7)

Removed

• Now explicitly excluding __pycache__ from releases by adding a prune statement to MANIFEST.in

3.1.2 - 2014-05-05

Added

• setup.cfg to support universal wheels (issue #6)
• Python 3.0 and Python 3.1 as requiring the argparse module

3.1.1 - 2014-03-01

Added

• Ability to sort lists of lists (issue #5)

Changed

• Cleaned up import statements

3.1.0 - 2014-01-20

Added

• signed and exp options to allow finer tuning of the sorting
• Doctests
• New shell script options that correspond to signed and exp
• In the shell script the user can now specify multiple numbers to exclude or multiple ranges

Changed

• Entire codebase now works for both Python 2 and Python 3 without needing to run 2to3
• Updated all doctests
• Further simplified the natsort base code by removing unneeded functions.
• Simplified documentation where possible
• Improved the shell script code
• Made the shell script documentation less “path”-centric to make it clear it is not just for sorting file paths

Removed

• The shell script filesystem-based options because these can be achieved better though a pipeline by which to filter

3.0.2 - 2013-10-01

Changed

• Made float, int, and digit searching algorithms all share the same base function
• Made the __version__ variable available when importing the module

Fixed

• Outdated comments

3.0.1 - 2013-08-15

Added

• Support for unicode strings (issue #2)

Fixed

• Empty string removal function

Removed

• Extraneous string2int function

3.0.0 - 2013-07-13

Added

• A number_type argument to the sorting functions to specify how liberal to be when deciding what a number is

Changed

• Reworked the documentation

2.2.0 - 2013-06-25

Added

• key attribute to natsorted and index_natsorted so that it mimics the functionality of the built-in sorted (issue #1)
• Tests to reflect the new functionality, as well as tests demonstrating how to get similar functionality using natsort_key

2.1.0 - 2012-12-05

Changed

• Reorganized package
• Now using a platform independent shell script generator (entry_points from distribute)
• Can now execute natsort from command line with python -m natsort as well

2.0.2 - 2012-11-30

Added

• The use_2to3 option to setup.py
• Include distribute_setup.py to the distribution
• Dependency to the argparse module (for python2.6)

2.0.1 - 2012-11-21

Added

• Tests into the natsort.py file iteself

Changed

• Reorganized directory structure

2.0.0 - 2012-11-16

Added

• Better README documentation
• Doctests

Changed

• Sorting algorithm to support floats (including exponentials) and basic version number support

Indices and tables

• Index
• Module Index
• Search Page