stephenb linux geek. foodie. libertarian. mobile addict.

Grep and Multiline Regular Expressions

Nothing new here: needed to go through some old log files to find something. Unfortunately, I needed to match across 2 lines specifically. I wrote up my regex, and tested it in Reggy and all was good.  I tried it out in the real world but didn't get any matches.  GNU Grep and Egrep don't support \n! It only searches one line at a time.  Enter: pcregrep -M.

(The other catch is that the files are compressed.  bzcat and zcat solve this problem.)

The entry I was looking for:

I 23:21:37:465  aaaaaaa  ZzzHandler: Subscription expired for x.
D 23:21:37:466  aaaaaaa  ZzzHandler: Proceeding to deactivate account

So I came up with this obfuscated 'for' loop:

for i in $(find /mnt/xxxxxxx -name xxxxxxx*bz2 -type f -print)
do
            echo $i; bzcat $i | \
            pcregrep -M ".*ZzzHandler: Subscription expired.*\n.*Proceeding to deactivate.*"
done | tee -a /tmp/xxxxxx-archived.log