The .index method does a linear search, checking on average 1/2 of the
items in the list. That's why it's so slow.
In order to avoid that you could build a dict of each value in
dimension_values[col_idx] and its index in a single pass so that it
becomes a quick lookup.
For example:
from itertools import count, izip
...
def main():
# dimension_names = ["color", "size"]
dimension_cols = [0, 1]
# measure_names = ["weight", "value"]
measure_cols = [2, 3]
indexers = [Indexer() for _ in dimension_cols]
facts = pairs(iterator_factory(),
dimension_cols, measure_cols, indexers)
facts = list(facts)
print facts
for i, indexer in enumerate(indexers):
print "%d: %s" % (i, indexer.to_list())
if __name__ == "__main__":
main()
Whew! :-) Thank you for taking the time. I'm not very familiar to
itertools yet, so I need some time to understand this *beautiful* code. :-)
L
--
http://mail.python.org/mailman/listinfo/python-list
