On Wed, Aug 1, 2018 at 9:56 PM, Richard Damon <rich...@damon-family.org> wrote: >> it says explicitly that numeric keys will use numeric comparison, but >> no mention is made of strings/bytes etc etc and there's an implication >> that object identity is used rather than comparison. In python 3.x >> b'a' is not the same as 'a' either the documentation is lacking some >> reference to strings/bytes or the warning is wrong. Using the excuse >> that normal comparison is being used seems a bit feeble. It would >> clearly improve speed if object identity were to be used, but I >> suppose that's not the case for other reasons. > > One problem with using identity for strings is that strings that have > the same value may not have the same identity. I believe very short > strings, like small numbers are cached, so same valued (and typed) short > strings will have the same identity, but longer strings are not. It > might work for 'a', but might not for 'employee'.
Strings and numbers MUST be compared by value, not identity. CPython has a cache for strings used in compilation, but that's about it: >>> s1 = "spam" >>> s2, s3 = "sp", "am" >>> s4 = s2 + s3 >>> s1 is s4, s1 == s4 (False, True) >>> n1 = 42 >>> n2 = n1 * 3 >>> n3 = n2 / 3 >>> n1 is n3, n1 == n3 (False, True) >>> n4 = 123456789 >>> n5 = n4 * 3 >>> n6 = n5 // 3 >>> n4 is n6, n4 == n6 (False, True) The middle example is guaranteed (because n1 is an int, n3 is a float), but the other two could in theory be cached by a compliant Python implementation. I don't know of any that do, though. ChrisA -- https://mail.python.org/mailman/listinfo/python-list
