One more thing I found while investigating #40, here's a bug I found which only occurs when fewMeasurements = "remove" and there are no rows that should be removed. This was a side effect of #40 , all what would have been NA were replaced with low numbers, so there were no missing values. See line 419 below
x <- x[-which(x$Feature %in% remove_feature_name$Feature), ]
When the which() statement returns an empty vector, you get x[-c(),] which selects no rows instead of the intended all rows except those inside c(). Recoding this with a boolean index instead of the -which() would fix this rare-to-occur bug.
x <- x[!(x$Feature %in% remove_feature_name$Feature), ]
One more thing I found while investigating #40, here's a bug I found which only occurs when fewMeasurements = "remove" and there are no rows that should be removed. This was a side effect of #40 , all what would have been NA were replaced with low numbers, so there were no missing values. See line 419 below
When the
which()
statement returns an empty vector, you getx[-c(),]
which selects no rows instead of the intended all rows except those inside c(). Recoding this with a boolean index instead of the -which() would fix this rare-to-occur bug.