Python 2

Use isinstance(obj, basestring) for an object-to-test obj.

Docs.

Python 3

In Python 3.x basestring is not available anymore, as str is the sole string type (with the semantics of Python 2.x's unicode).

So the check in Python 3.x is just:

isinstance(obj_to_test, str)

This follows the fix of the official 2to3 conversion tool: converting basestring to str.

Python 2

To check if an object o is a string type of a subclass of a string type:

isinstance(o, basestring)

because both str and unicode are subclasses of basestring.

To check if the type of o is exactly str:

type(o) is str

To check if o is an instance of str or any subclass of str:

isinstance(o, str)

The above also work for Unicode strings if you replace str with unicode.

However, you may not need to do explicit type checking at all. "Duck typing" may fit your needs. See http://docs.python.org/glossary.html#term-duck-typing.

See also What’s the canonical way to check for type in python?

Python 2 and 3

(cross-compatible)

If you want to check with no regard for Python version (2.x vs 3.x), use six (PyPI) and its string_types attribute:

import six

if isinstance(obj, six.string_types):
    print('obj is a string!')

Within six (a very light-weight single-file module), it's simply doing this:

import sys
PY3 = sys.version_info[0] == 3

if PY3:
    string_types = str
else:
    string_types = basestring

I found this ans more pythonic:

if type(aObject) is str:
    #do your stuff here
    pass

since type objects are singleton, is can be used to do the compare the object to the str type

If one wants to stay away from explicit type-checking (and there are good reasons to stay away from it), probably the safest part of the string protocol to check is:

str(maybe_string) == maybe_string

It won't iterate through an iterable or iterator, it won't call a list-of-strings a string and it correctly detects a stringlike as a string.

Of course there are drawbacks. For example, str(maybe_string) may be a heavy calculation. As so often, the answer is it depends.

EDIT: As @Tcll points out in the comments, the question actually asks for a way to detect both unicode strings and bytestrings. On Python 2 this answer will fail with an exception for unicode strings that contain non-ASCII characters, and on Python 3 it will return False for all bytestrings.

Its simple, use the following code (we assume the object mentioned to be obj)-

if type(obj) == str:
    print('It is a string')
else:
    print('It is not a string.')

In order to check if your variable is something you could go like:

s='Hello World'
if isinstance(s,str):
#do something here,

The output of isistance will give you a boolean True or False value so you can adjust accordingly. You can check the expected acronym of your value by initially using: type(s) This will return you type 'str' so you can use it in the isistance function.

I might deal with this in the duck-typing style, like others mention. How do I know a string is really a string? well, obviously by converting it to a string!

def myfunc(word):
    word = unicode(word)
    ...

If the arg is already a string or unicode type, real_word will hold its value unmodified. If the object passed implements a __unicode__ method, that is used to get its unicode representation. If the object passed cannot be used as a string, the unicode builtin raises an exception.

isinstance(your_object, basestring)

will be True if your object is indeed a string-type. 'str' is reserved word.

my apologies, the correct answer is using 'basestring' instead of 'str' in order of it to include unicode strings as well - as been noted above by one of the other responders.

This evening I ran into a situation in which I thought I was going to have to check against the str type, but it turned out I did not.

My approach to solving the problem will probably work in many situations, so I offer it below in case others reading this question are interested (Python 3 only).

# NOTE: fields is an object that COULD be any number of things, including:
# - a single string-like object
# - a string-like object that needs to be converted to a sequence of 
# string-like objects at some separator, sep
# - a sequence of string-like objects
def getfields(*fields, sep=' ', validator=lambda f: True):
    '''Take a field sequence definition and yield from a validated
     field sequence. Accepts a string, a string with separators, 
     or a sequence of strings'''
    if fields:
        try:
            # single unpack in the case of a single argument
            fieldseq, = fields
            try:
                # convert to string sequence if string
                fieldseq = fieldseq.split(sep)
            except AttributeError:
                # not a string; assume other iterable
                pass
        except ValueError:
            # not a single argument and not a string
            fieldseq = fields
        invalid_fields = [field for field in fieldseq if not validator(field)]
        if invalid_fields:
            raise ValueError('One or more field names is invalid:\n'
                             '{!r}'.format(invalid_fields))
    else:
        raise ValueError('No fields were provided')
    try:
        yield from fieldseq
    except TypeError as e:
        raise ValueError('Single field argument must be a string'
                         'or an interable') from e

Some tests:

from . import getfields

def test_getfields_novalidation():
    result = ['a', 'b']
    assert list(getfields('a b')) == result
    assert list(getfields('a,b', sep=',')) == result
    assert list(getfields('a', 'b')) == result
    assert list(getfields(['a', 'b'])) == result

You can test it by concatenating with an empty string:

def is_string(s):
  try:
    s += ''
  except:
    return False
  return True

Edit:

Correcting my answer after comments pointing out that this fails with lists

def is_string(s):
  return isinstance(s, basestring)