kevinblighe / RegParallel

Standard regression functions in R enabled for parallel processing over large data-frames.
37 stars 12 forks source link

One/few models failing lead to whole job failing #14

Open gatorchris4 opened 1 month ago

gatorchris4 commented 1 month ago

I am using the RegParallel package to conduct a coxph on many SNPs (hundreds of thousands). I am getting errors in a few of my datasets, not all. I know the error is because some SNPs don't have enough observations in either the outcome or some of the covariates. Instead of filtering out SNPs, I was wondering if there is a way for the RegParallel function to "skip" variables that lead to an error but still process the rest of them that wouldn't lead to an error because I don't want to systematically remove SNPs since each might fail for different reasons (either outcome or covariates with not enough observations).