Identify INDICES value that 'breaks' by() [or, equivalently, tapply()]

Question

When using the function by, at times I will have a data subset (as determined by the INDICES argument) that 'breaks' by (technically it breaks FUN which in turn breaks by).

Is there a way to identify the 'bad' value of the list passed to INDICES? (without writing an explicit loop over the list)

Answer 1

Can you make a reproducible example?

You can try something like,

customfun <- function(x)try(somefun(x))
byresult <- by(mydata, mydata$group, FUN=customfun)
whichbad <- sapply(byresult, function(x)which(inherits(x, "try-error")))