factors-memory.R

# Factors are more memory efficient (if labels > few bytes), since redundant multi-byte
# labels are stored once in memory (as attributes), and integers keep the mapping. E.g.:

a = sample(paste0("chrom", c(1:22, "X", "Y")), 1e8, replace=TRUE)
object.size(a)
# 800001192 bytes
object.size(factor(a))
# 400001744 bytes

# For long character vectors of repeating values, this *really* pays off.