You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I suspect that the matrix multiply for a dgCMatrix %*% IterableMatrix does not include the names for at least one of the dimensions. Does that make sense? Here is what I see
The R code looks like
mapping <- my.dMcast(groupings2, form)
colnames(mapping) <- substring(colnames(mapping), 2)
result <- Matrix::t(mapping) %*% x
# result has no colnames
browser()
Thanks for the detailed example code, which helped a lot in diagnosing the issue.
The core problem was due to some weird corner case bug I had in convert_matrix_type() -- basically when two convert_matrix_type() calls went from double -> uint32_t -> double, then the dimnames were lost on the second call. The binarize followed by matrix multiplication was triggering this pattern of conversions internally
I've fixed this in the main branch now (though please comment if my fix didn't actually solve your issue)
Hi,
I suspect that the matrix multiply for a dgCMatrix %*% IterableMatrix does not include the names for at least one of the dimensions. Does that make sense? Here is what I see
The R code looks like
and the mapping, x, and result objects look like
I appreciate your consideration and guidance.
Thank you.
Ever grateful,
Brent
The text was updated successfully, but these errors were encountered: