I definitely agree with this. Most of the files I am working with would give me the "out of memory" message if I tried to slurp then entirely into memory. They are just way too large. Processing one line at a time is quite often the solution. But I can't say more without seeing the program and a data sample.