How to replace ALL NA in ffdf columns?

Question

I have an ffdf object with 200K rows and 12K columns.

This is from a LEFT JOIN with another ffdf object, so has a lot of NA across the different columns.

How can I replace all NA with a specific value (say, FALSE, since the additional columns from the LEFT JOIN are all LOGICAL)?

I can successfully use the following syntax over one column to replace all NA in that column ("coltest"):

ffdfOut$coltest <- with(ffdfOut, ifelse(is.na(coltest), FALSE, coltest))

but the following loop:

cnamesLogical) <- colnames(ffdfOut)[12:12000]
for(colname in cnamesLogical)
{
 ffdfOut[, colname] <- with(ffdfOut, ifelse(is.na(colname), FALSE, colname))
}

gives me error:

Error in with.ffdf(ffdfOut, ifelse(is.na(colname), FALSE, colname)) : 'with.ffdf' only returns ff object of equal length of nrow(data)

Thanks for any help on this!

Answer 1

您可以使用lapply()遍历数据集的每一列

lapply(ffdfOut[,12:12000], FUN=function(coltest) ifelse(is.na(coltest), FALSE, coltest))