I have a 50GB file with a JSON string record per file, that has a very high variance in record length. So I do not want to split it based on number of lines, but based on the size of each fragment. However, when I use:
split -n 100000000 .json
where each split file should be roughly 100 MB.
I find that trailing and leading records are sometimes truncated in my xaa, xab, xac.... etc files. Do you know how I can solve this problem?