"sofeng" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED] |I would like to use the following recipe to transpose a list of lists | with different lengths. http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/410687 | | Here is an example of what I would like to do: | | Python 2.5.1 (r251:54863, May 18 2007, 16:56:43) | [GCC 3.4.4 (cygming special, gdc 0.12, using dmd 0.125)] on cygwin | Type "help", "copyright", "credits" or "license" for more information. | >>> a = [[1,2,3],[4,5,6,7],[8,9]] | >>> print map(lambda *row: list(row), *a) | [[1, 4, 8], [2, 5, 9], [3, 6, None], [None, 7, None]] | >>> | | However, in the Python 3000 FAQ (http://www.artima.com/weblogs/ | viewpost.jsp?thread=211200), Guido says not to use map with lambda | because a list comprehension is clearer and faster. | | How can I rewrite the above recipe using a list comprehension instead?
Here is the sort of thing Guido is talking about: >>> a = [[1,2,3],[4,5,6,7],[8,9]] >>> [list(r) for r in zip(*a)] [[1, 4, 8], [2, 5, 9]] Except in this case, zip does not pad, while map does, so the list comp form has to be >>> [list(r) for r in map(None, *a)] [[1, 4, 8], [2, 5, 9], [3, 6, None], [None, 7, None]] but at this point, it is about as easy to put the lambda in place of None. But when one maps one sequence or equal-length sequences, Guido's point applies. tjr -- http://mail.python.org/mailman/listinfo/python-list
