Alex, I am using Numeric and have created 3 arrays: zero((1215,1215),Float) Two arrays are compared and one is used to hold the mean difference between the two compared arrays. Then I compare 290 or 340 pairs of arrays. I know that memory is a problem and that is why I don't open all of these arrays at the same time. I cannot install Numpy due to my working conditions. Sorry I should have made it clear that is was Numeric I was working with.
/Sheldon Alex Martelli wrote: > Sheldon <[EMAIL PROTECTED]> wrote: > > > Hi Gary, > > > > I am really trying to cut the time down as I have 600+ arrays with > > dimensions (1215,1215) to compare and I do a lot more things with the > > arrays. If I understand you correctly, there is no way around a for > > loop? > > In pure Python (w/o extension packages) there are no 2-D arrays; so > either you're using lists of lists (and I wonder how you fit even one of > them in memory, if they're 1215 by 1215, much less 600!) or you're > already using some extension (Numeric, numarray, numpy) and aren't > telling us which one. If you're using pure Python add your extension of > choice, if you're using an extension already tell us which one, and in > each case there will be ways to perform your manipulation tasks faster > than Python-level for loops would afford. > > > Alex -- http://mail.python.org/mailman/listinfo/python-list