Closed marshallaf closed 7 years ago
I've pushed a fix that keeps the dimensions from being dropped when there is just one row. See if it works for you. It should be in the trunk branch here on GitHub.
Thanks! That fixed that particular issue. And much more concisely than my approach!
I discovered that my real problem that I had was that I had a rangeGate
earlier in the workflow that had a too-high refline, so I was calling proBin
with minEvents = 0.45
, which even with this fix gave me some weird stats for the bins.
I'm getting an error while using
proBin
on certain fcs files:dim(X) must have a positive length
It occurs in this call to
apply
, and, as far as I can tell, happens when the data at that node has only 1 row.The selection made in the
apply
call -data[[as.character(nodeIndx)]][,channels]
- returns a vector, whichapply
doesn't work on sincedim(vector) == NULL
.My grasp on R style and syntax is still pretty lacking, but I think if the condition was checked for, then the vector could be turned into a matrix by:
Although there's probably a better way of doing that.