Text search ..

[dbeez]

I'm looking for a simple command that will run down a directory structure and print out the name of any file containing a certain string.

I've tried using find, but it doesn't want to work and keeps giving me errors.

What is the easiest way to do this ??

Thanks

 

[PHV]

it doesn't want to work and keeps giving me errors
Any chance you could post the error messages and the code you tried ?

Hope This Helps, PH.
Want to get great answers to your Tek-Tips questions? Have a look at FAQ219-2884 or FAQ181-2886

 

[dbeez]

Hi ph,

I think I've got that bit now, but I have a new problem (of course!). My code this time looks like this

if [ (grep string `find . -print`) != 0 ]
then <search for email address>
<print email address to file>
fi

... so what I want to do basically is run down the directory structure and find pages with string in them, then to write the email addresses contained in those pages to a file.

... statement at the top is obviously messed up, but I don't know what kinda return statement to expect from a grep that is unsuccessful. I'm guessing that it would be 'exit 0'

any help here is gratefully accepted

 

[PHV]

You may try something like this:
find . -type f | while read f
do grep -q 'string' "$f" && grep '@' "$f"
done > /path/to/output

Hope This Helps, PH.
Want to get great answers to your Tek-Tips questions? Have a look at FAQ219-2884 or FAQ181-2886

 

[dbeez]

thanks phv,

that helps enormously with scrolling through the files. I'm still trying to pick out the addresses though.
So far I've got

find . -type f | while read f
do grep -q 'string' "$f" && sed 's/.*mailto://;s/".*//' "$f" | grep '@'
done > file.email

... this is about as close as I've gotten. Unfortunately though it still gives me lines with @ and no email address.

... so what I'm doing here is looking for the email address wrappers (eg. mailto and then excluding them. What I really want to do is look for the email address and then exclude everything else.

Can't seem to be able to do that though. Frankly I'm finding man sed a little confusing at the moment.

thanks again ...

 

[feherke]

Hi

Frankly I'm finding man sed a little confusing at the moment.

Right. And incomplete too. At least my GNU [tt]sed[/tt]'s man page does not mention all features. This is a better documentation :

http://www.gnu.org/software/sed/manual/html_mono/sed.html

Feherke.

http://rootshell.be/~feherke/

 

[Annihilannic]

find . -type f -print0 | xargs -0 grep -i 'mailto:.*string' | sed 's/\([^:]*\):.*mailto:\([^" >]*\).*/\1: \2/i'

Or, if you have a grep that supports -r for recursive searches:

grep -ri 'mailto:.*string' . | sed 's/\([^:]*\):.*mailto:\([^" >]*\).*/\1: \2/i'

Breaking down the sed command:

s/ search and replace

\([^:]*\) all the non-colon characters before the first colon (i.e. the filename)

.*mailto: the line up to and include the mailto: component (to be discarded)

\([^" >]*\) the email address up to the first space, speech mark or tag terminator

.* the rest of the line (to be discarded)

/

\1: \2 replace it with the filename \1 and the matching email address \2. \1 and \2 are derived from the bracketed components of the search string.

/i and ignore case.

Annihilannic.

 

[dbeez]

X-cellent ... thanks feherke

I had most of this script done a couple of weeks ago but lost it when I reinstalled and forgot to save it.

I'll read over the reference in the morning ... I may still have a few questions later on though

thanks again guys ...

 

[dbeez]

thanks anni*

I can't seem to get your code to work unfortunately, it just spits out

root@ubuntu:/home/babo/Desktop/spider_proj # grep -ri 'mailto:.*string' . | sed 's/\([^:]*\):.*mailto:\([^" >]*\).*/\1: \2/i'
./spider: \([^
./spider: \([^

... also, I'm looking to parse a file that matches my string, as opposed to a line that matches my string.

... thanks for explanation though, even though I do have one question. About the

\1: \2 replace it with the filename \1 and the matching email address \2.  \1 and \2 are derived from the bracketed components of the search string.

... I realize that it is derived from the brackets ... but what exactly is it meant to do ???

... also what exactly does the -e switch do ... I realize that it includes a script file in the command, but why do people use it even when they aren't using a script file ???

I realize that these are probably really dumb questions - so sorry in advance

 

[dbeez]

Ok cool, I think I've found something a good text on sed ...

Feherke I think your suggestion was missing quite of bit of the sed command as well. The best one I've come up with is probably

http://www.grymoire.com/Unix/Sed.html

... hopefully I'll be able to take it myself from here ... thanks guys

 

[Annihilannic]

dbeez,

I think it worked, however it found a matching string in ./spider, which of course was part of the sed command, not an email address.

The output of my commands should be something like:

[tt]./filename: email@address.one
./subdir/filename2: email@address.two[/tt]

\1, \2, \... are used for pulling chunks out of the matched
string and putting them into the replacement. Some more examples:

s/.*\([0-9]\)*.*/\1/ would pull the first occurrence of a number out of a line.

s/\([a-z]\) \([a-z]\)/\2 \1/ would swap around two lower-case strings, e.g. "apples oranges" would become "oranges apples".

These work in vi too of course.

Regarding -e, maybe this chunk of man page (from Solaris) will clarify:

[tt] -e script
script is an edit command for sed. See USAGE below for
more information on the format of script. If there is
just one -e option and no -f options, the flag -e may
be omitted.[/tt]

I'll leave adapting my solutions to your requirements as an exercise for you since you seem to be well on the way there.

Annihilannic.