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 don't think this affects the users, but I found out this issue when writing a cbind method for a class that inherits from SCE and it caused a unit test to fail. Here's the reprex:
# From an example from the SCE package
ncells <- 100
u <- matrix(rpois(20000, 5), ncol=ncells)
v <- log2(u + 1)
pca <- matrix(runif(ncells*5), ncells)
tsne <- matrix(rnorm(ncells*2), ncells)
sce <- SingleCellExperiment(assays=list(counts=u, logcounts=v),
reducedDims=SimpleList(PCA=pca, tSNE=tsne))
thanks for reporting this. @LTLA is this the intended behavior?
I would imagine that each object should have only one version, but on the other hand, can you combine two objects with different versions? And in this case, would you want to keep track of this? Or which version should "win"?
Probably just a consequence of mimicking the default behavior for SE's cbind.
In this case, the correct behavior would be to replace both versions with the current version, given that cbind should return an SCE with its internal structure updated to the current version of the package.
If both are replaced by the current version, then there really should be one version field in the int_metadata in the concatenated object, not two duplicated ones.
I don't think this affects the users, but I found out this issue when writing a
cbind
method for a class that inherits from SCE and it caused a unit test to fail. Here's the reprex:Then see the
int_metadata
in the R console.Now I concatenate the same SCE object
And got this:
The text was updated successfully, but these errors were encountered: