delete files based on keywords
 

How do I delete files based on a keyword.

Lets say in a directory I have thousands of files and I want to delete those that have "cialis" word on it, how would I do that?

You could do something like find / *cialis* | xarg rm -f

I might have the expressions wrong. Someone correct it?

Or, you could just go into that directory and do ls *cialis*. If the files show up correctly, you can just do rm *cialis*

Have you considered something like SpamAssassin ? since it sounds like you are trying to filter spam. from a MailDir


You may find this resource useful as well since they never spell it Cialis or Viagra without trying to disguise it.. http://wiki.castlecops.com/A_list_of_Regex_topics

the word "cialis" is inside the files


you're right, im dealing with emails inside Maildir, We do have spamassassin but it is disabled at the moment.

if i do this
ls | xargs grep viagra |more

it will have a result of like this
1182027917.M532957P11494V0000000000000808I0007E23D_0.server.com,S=4542:2,: CIALIS (super viagra) FOR AS LOW AS $4.38 PER DOSE
1182027917.M532957P11494V0000000000000808I0007E23D_0.server.com,S=4542:2,: <strong>CIALIS</strong> (super viagra) FOR AS LOW AS <strong>$4.38
</stron=
1182097726.M794748P29030V0000000000000808I0011988B_0.server.com,S=6068:2,:Activities ngos capacity effective, starts small. Free lesbians, xxx
air fares generic viagra adipex plavix. Pages go, section, choose national. Hurdle number lofty legal.
1182097726.M794748P29030V0000000000000808I0011988B_0.server.com,S=6068:2,:lesbians, xxx air fares generic viagra adipex plavix. Pages go, sect
ion, =

so what's the next step, I want to delete that result?

should do the trick... untested, try with echo instead of rm first ;}


Cheers,
Tink

Assuming you are in the directory you want to search:

find . -exec grep -l 'cialis' "{}" \;

You could then

find . -exec grep -l 'cialis' "{}" \;|xargs rm

Did not see post above. Practically the same

Remember that grep returns TRUE if it has found a match and false otherwise.

You could use
grep '<pattern>' "$file" && rm "$file"
inside a loop.

Since you may be looking for multiple patterns, you could use a file containing the patterns to look for:

patterns:
[vV]iagra
[cC]ialis

and then use
grep -f patterns "$file" && rm "$file"
inside a loop. Now you can simply edit the patterns file to make changes or add additional patterns.

thanks for the replies, guys.

ill try those examples and see what happens, thanks again.

grep -l 'cialis' will look for cialis and words with cialis like specialist, how can make it search for the exact match?

add space on each side eg
grep ' cialis ' $file

very nice, it works great


another question

how can I see the files being deleted when executing
find . -exec grep -l ' meds ' "{}" \; | xargs rm

Use "tee" to write find's output in a file. This will give you a list of files. Alternatively, you can use rm's "-v" option.

Please change "rm" with "echo" or "ls" the first time you run it, just to avoid deleting the wrong files.

Examples:

timmeke
i just tried rm -v before you posted your answer :D, but thanks anyway those other examples helps :)

a follow-up question...

using grep -f <file> will process all patterns in that file, will it search for the exact pattern or not?