On Aug 16, 11:02 pm, "Carsten Haese" <[EMAIL PROTECTED]> wrote: > On Fri, 17 Aug 2007 03:15:10 -0000, beginner wrote > > > On Aug 16, 9:32 pm, Carsten Haese <[EMAIL PROTECTED]> wrote: > > > What is the best solution in Perl need not be the best solution in > > > Python. In Python you should just use a tuple as your dict key, i.e. > > > a[k1,k2] = v, unless you have some other constraints you're not telling > > > us. > > > > HTH, > > > > -- > > > Carsten Haesehttp://informixdb.sourceforge.net > > > I use tuples this way all the time. It is indeed very neat. But it > > is not a replacement for double hash-table. If I want to retrieve > > information just by K1, it is not efficient to index on (K1, K2). > > If you have to look up all values associates with k1 and any k2, you're right, > that's not efficient. That would fall under "other constraints you're not > telling us." I'm not a mind reader. > > -Carsten
Yeah, I should have mentioned that I actually want to group the data by K1 and then by K2. -- http://mail.python.org/mailman/listinfo/python-list