Hello,
I've got a program that creates its own log file in normal text file, what i would like to be able to do is create another bash script that could perhaps sort through the log file and pick out important lines of it depending on the test string contained on the line. Rather like how swatch works but i'd like my own independent script. Its not something that will run constantly or on a loop just something that the user can run every now and then and it pulls in the current log file and re-creates it only showing specific strings i've listed in another file for example.

does anyone know of a script already out there that could customise to my own needs, or will i need to start from scratch, probably using grep's etc etc.

Tanks for any guidance

Use grep to search the file for specific strings.

Personally I'd use perl - maybe have a look at logwatch.

Sounds like a job for Bash and Awk. If you have interest in a number of log files (1 or more), identify them in a selection file and in a loop have Awk create/append to your own chosen extract file. Awk is a super choice for selecting from your logs depending on pattern matches and formatting the output. When your loop of 1 or more files completes, head up and tail the extract with the appropriate email information and pipe the lot into xmail or what-have-you. after which you can either archive or delete your extract file ready for the next run. If you want you could have it run each day or once a week etc using at and just have it arrive in time for reading with your coffee. The possibilities are fairly endless and could save a lot of trawling through logs if you know what it is that you are interested in and if the logs have unique tags to allow particular types of event to be pulled out easily. Happy hacking.

PAix

yep thats all rather along the lines what i was thinking, sorry haven't really taken a look at perl yet (not had the need to), and logwatch i also know of but don't require. But i didn't know of a decent way to script it. I suppose its a grep issue i have, i'd like to read in each line of text output in a text file, and if a string of text on a line is a match of any of the strings of text i have in another file that i'd like the text to be compared against, then for it to consider that as a match for that line of text copy it to another file and then check the next line of text for other matches from the other file, if a match again print line to other file.

Something like / similar to this would be amazing but not sure if grep can handle it, might need the use of sed as well???

Can you please provide an example of your input and expected output.

something like log file might contain:
thats an example of log file obvious not the exact one but still you can get the idea.
Now a file of text strings to match against.
so if any string of text on a line in log.txt matches any of the text strings from matchtext.txt, i'd like to be able to take the line thats matched and copy paste so to speak, probably more a redirect to another output file (another log file) called sivedog.txt

Thanks regards

awk sounds like a good way of accomplishing this with a script (easier than sed anyway).

does anyone have a script, or can anyone help me write one or start an initial outline of one for this problem??

Thannks and Regards

Use nested loops or wildcard for log files if more than one

Thats fantastic thanks!!!!
Worked like a charm

any time