Execution halted for geneCounts() on c_call

[Scott Christley]

When I run countGenes() on c_call where the field doesn’t have any data, I get an execution halted, which kills the whole batch processing of the file. Is there any way for this to gracefully fail? E.g., return count with NA value, or not abort the script so it can continue running?

> genes <- countGenes(db, gene='c_call', group='repertoire_id', mode='allele')
Error in countGenes(db, gene = "c_call", group = "repertoire_id", mode = "allele") : 
  The column c_call contains no data
Execution halted

[Jason Vander Heiden]

Hrm. What would be the ideal output in this case? And can you get there via a tryCatch? Eg:

genes <- tryCatch(countGenes(ExampleDb, gene="c_call", mode="asis"), error=function(e) NA)

Also, it looks like the allele regex isn’t working for C regions right now (IGHD specifically), so use mode="asis" with c_call. I’ll open a separate issue to fix that.

[Scott Christley]

I think for the other gene calls, maybe specifically d_call, I get a NA row. I’m guessing those are the rows which have a blank d_call

Ok, I’ll try the tryCatch. Do the gene and familymodes made sense for c_call?

[Jason Vander Heiden]

The gene and family modes don’t work correctly either, for the same reason.

[Jason Vander Heiden]

Oh, I don’t think you should get NA for a row… If the gene is missing in a given group then the row for that gene should either be missing (default; fill=FALSE) or assigned a count of 0 (fill=TRUE). If you’re getting NA rows, then I suspect there’s a bug.

[Scott Christley]

I thought maybe because using readChangeoDb() was causing the NA row, but I get it with airr::read_rearrangement as well. I’ll file a separate issue and provide the input file.