Op 2004-12-13, Fredrik Lundh schreef <[EMAIL PROTECTED]>: > "jfj" wrote: > >> Why can't we __setitem__ for tuples? > > http://www.python.org/doc/faq/general.html#why-are-there-separate-tuple-and-list-data-types > >> The way I see it is that if we enable __setitem__ for tuples there >> doesn't seem to be any performance penalty if the users don't use it >> (aka, python performance independent of tuple mutability). > > how would you implement a dictionary where the keys could change, without > any performance penalty compared to the current implementation?
The performace gained by using tuples as keys in dictionaries is entirely illusional. Sure the fact that you use a tuple which is immutable, makes that you can put the key directly in the dictionary instead of a copy and that will gain you some performance. But this performance gain can be more than offset by other code in the program. Suppose you need a list/tuple as a key and most opperation you will do on those keys will be appends and pops. You now have the itwo choices 1) Always convert your lists to tuples on key entries and keys accesses, which will mean more copying than when a copy of a key would have been made on key entry. 2) Simulate appends and pops by tuple operations which can also require more copying than was gained by using tuples as key. -- Antoon Pardon -- http://mail.python.org/mailman/listinfo/python-list
