grep regular expression
 

i work in finance.

i was looking through a 80 meg file to verify if accounts are attached to a process. some processes can have up to 10,000 accounts attached to them, some can have just one.

i tried this :

grep cdu115 ~/data/flat_files/account.t

and came up with thousands of accounts attached to processes cdu115a, cdu115c, cdu115f, and so on- but what i was looking for was just plain old cdu115. even if just one account is attached to this process, it is valid.

so i tired this

egrep '^cdu115$' ~/data/flat_files/account.t - which resulted in nothing, so the process is invalid. but just to be sure i tried:


egrep '^cdu115a$' ~/data/flat_files/account.t - which also resulted in nothing - but is false because when i greped cdu115 from this file i saw thousands of entries for cdu115a.

so i am uncertain.

I tried grep -e, double quotes, encasing the regex in / ... / .

i need to be certain that there are no accounts attached to process cdu115 before i delete it. and that is cdu115 without extremeties - no cdu115a, cdu115b, cdu115c, cdu115d, ect.

how can i be certain in regex terms that i am looking for just cdu115 and not cdu115a, cdu115b, cdu115c, ect.

what is the proper regex for this situation?

You need to read the manpage (better) to see how to limit the match to "boundaries". Several ways to do it with grep - easiest in your case is probably to use the "-w" switch.

Welcome to LQ!!

There's also the "word boundary" regex....eg:

grep '\<cdu115\>' filename

This works in utilities other than grep.