Hello, Recently, I needed to grep multiple files in a directory, using 28,212 matching patterns. I did this two ways. The first way, I had all 28,212 patterns in one file, (say filename, and used
grep -f filename files > results1 The second way, I split the one large file into 28 or so smaller files (say filename1, filename2, and so on...) and used grep -f filename1 files > results2 grep -f filename2 files >> results2 ...and so on...basically iterating 28 or so times over files. By the time each process finished, for the first process (28,212 arguments in one file) I had a results file that was 1.6MB. But for the second process (28 files,1000 or so arguments per file, iterating grep once for each file ), I ended up with a 5.9MB file. I wonder if this is enough to assume that grep has trouble with too many arguments read in from a file? If that is so, then a safe limit would be somewhere between 1000 and 28,211. Is there a consensus out there as to whether or not there's a point at which greping so many arguments in from a file shouldn't be done? Thanks! Tony Zanella -- View this message in context: http://www.nabble.com/best-practices-for-grep--f--tp20835034p20835034.html Sent from the Gnu - Grep mailing list archive at Nabble.com.
