[Numpy-discussion] Adapting algorithm to accept Scipy sparse matrix
Wed Mar 21 10:04:33 CDT 2007
> The "worst" problem I encountered is that sparse matrices do not
> seem to support the kind of indexing I need. At least I get
> "NotImplementedError: sequence indexing not yet fully supported"
> and "<sparse matrix type> supports slices only of a single row"
> errors all the time.
I agree .. this is kind of sucky. There was a (short) thread (I
think) a few weeks ago that had mentioned something about doing more
work on the sparse matrix stuff, but I can't seem to find it.
The day when we'll have a unified interface to normal/dense matrices
and sparse matrices will be a great day indeed ... well, ok, at least
a pretty darn good day ...
More information about the Numpy-discussion