When fread has to bump many columns to a different data type, R can crash. This dataset with 10k rows and 40 columns makes R consistently crash:
library(data.table)
# Create a character column that is mostly empty,
# so it gets bumped.
x <- character(1e4)
x[9994] <- "+"
# For very powerful machines, you might need to
# increase the number of rows a little.
dfr <- as.data.frame(
replicate(40, x, simplify = FALSE)
)
write.table(dfr, "test.txt", quote = FALSE)
fread("test.txt")
When fread has to bump many columns to a different data type, R can crash. This dataset with 10k rows and 40 columns makes R consistently crash: