Shell Programming and Scripting

can someone please show me a better way (if there is one), to grep out all variations (upper and small case letters) from a file?

Code:

egrep "(panic|error|failed|fail|FAIL)" /var/log/messages

I'm interested in finding any string that may indicate somethings wrong with ANY part of a system. so, if you know of any other strings outside of the ones specified above, please please please feel free to add them to the code you provide.

my systems run linux and sunos oses.

thanks

What's WRONG for one may not be WRONG for another. Anyways, you place the WRONG keywords in a file (say, patts) and then do a case-insensitive grep using this pattern file...

Contents of patts could be :

Code:

FAIL
PANIC
ERROR

Then, you could use

Code:

grep -if patts <filename>

Or similar to what you had already:

Code:

egrep -i "panic|error|fail" /var/log/messages

Like in elixir sinari's example, the -i switch is for "ignore case".

i was looking for something more complex. the "-i" will be slower when grepping through large files, which is what i'm dealing with.

i was thinking something along the lines of:

Code:

egrep "([pP]anic|[Ee]rror|[Ff]ail)" /var/log/messages

in this example, i know which letters may be small/upper case. so i took that into consideration. there are cases when u wont know, and i wanted to find a way to account for that, without using the "-i".

At any rate leave out the grouping operators ( ... ) , they seem to really slow things down...

oh really? i wasn't aware of that all!

It probably depends on implementation, so you would need to find out if they matter in your case, but FWIW, a couple of times I found these to be the culprit. Perhaps because grep not only needs to create groups, but a back reference as well. In any case you do not need them here...