In Python both single ' '
and double " "
quotes can be used to enclose the characters in a string.
For conventional strings by default, single quotations should be used, making your code more consistent with Pythons documentation. For example:
chr(97)
Returns:
'a'
When the string contains a string literal, double quotations should be used. For example "the str is 'hello'"
.
The triple single quotes are used for a multi-line comment:
text = '''Takes a singular word
and returns its plural.
'''
If the multi-line string contains a string literal, triple double quotations should be used:
text = """Takes a singular word
and returns its plural.
e.g. 'apple' becomes 'apples'.
"""
That being said PEP 257 recommends always using triple double quotations """ """
for docstrings for consistency. The reason it recommends using triple double quotations, even if it is a single line docstring is it makes it easier to expand the docstring into multiple lines later on.
Likewise because docstrings themselves are more likely to contain string literals, using triple double quotations also makes it easier to expand later on. The simple example below shows a one line docstring:
def plural(word, /):
assert isinstance(word, str)
"""Takes a singular word and returns its plural."""
return word + 's'
This can later be expanded:
def plural(word, /):
assert isinstance(word, str)
"""Takes a singular word and returns its plural.
e.g. 'apple' becomes 'apples'.
Parameters
----------
word : str
singular str.
Returns
-------
str
plural str.
"""
return word + 's'
Raymond Hettinger a Python core developer has made the argument that single quotations should be used where possible as the language itself prefers using single quotations. He has however noticed that the Python community itself seems to move towards a preference for use of double quotations.
He has noted that this is likely due to programmers who have migrated from languages such as C++ which use single quotes for a character 'a'
and double quotes for a string "apples"
. I initially came from MATLAB to Python for data science and also initially tended to favour double quotations for the same reason. This habit was reinforced as most of the data science tutorials I followed also favoured double quotations. It is better practice to be consistent with the Python language itself where possible, so I now follow Raymond Hettinger's recommendation to use single quotations for a regular string, double quotations for a string containing a string literal and use triple double quotations for a docstring.