Shell Programming and Scripting

Hello everyone,

I'm a newbie here. I'm working on a ksh script on Solaris 5.10 and have a question. I'm trying to search for a word line by line in a log file, and grab the text right after it inside quotes, and assign that to a variable.

For example, each line in the log will look something like this:
TERMINAL: "blahblah1" USERID: "blahblah2" USERHOST: "blahblah3"

I want to search for "USERID:" and assign the text in quotes right after it to another variable - so this new variable will = blahblah2.

Can anyone assist? I've been reading about sed, awk, regular expressions, and can't figure it out.

Thanks so much!

Try this, the text between the quotes will be in $uid:

Code:

perl -ne '/USERID:\s"(.+?)"/; print $1, "\n"' logfile | while read uid
do
   # your stuff here
done

I should have added that I'm going to be doing this for about 10 different strings per line. The code I have so far is:

Code:

cat $<logfile> | while read a
do
variable1=`echo $a | grep "USERID:" ???? `
#variable2=
#variable3=
#copy these values to another file separated by commas
done

Code:

USERID=Kristin_in_CO

awk -v u=$USERID '{if ($3==u ":") {gsub(/"/,"",$4);print $4}}' logfile

I ended up using what pludi suggested and it works perfectly. Thanks for your help!

cat $<logfile> | while read a
do
userid=`echo $a | perl -ne '/USERID:\s"(.+?)"/; print $1, "\n"'`
#do more stuff
done

So I ran my script with the suggested perl statement inside the while loop, and even though it completed successfully, it took over 6 hours to run reading line by line in a 90MB log file. I've decided to extract each of the 10 columns out individually like what's below, write them to their own file, and then paste each file (line by line) together into one log file. Doing it that way only took 1 minute to run.

Anyway, this perl statement doesn't seem to work outside of the while loop when there is a null value.

perl -ne '/USERID:\s"(.+?)"/; print $1, "\n"' datafile.log > userid.log

Example from datafile.log:
<null>
<null>
kristin
<null>
<null>
kristin
.......

What shows up in the userid.log:
<null>
<null>
kristin
kristin
kristin
kristin

I don't know why the 4th and 5th value don't show up as null in userid.log. I've been researching all day and trying different things but can't seem to get past the null issue. I also tried that suggested awk statement and got these errors:
awk: syntax error near line 1
awk: bailing out near line 1

Any help is much appreciated!

---------- Post updated at 06:43 PM ---------- Previous update was at 06:16 PM ----------

Sorry forgot to add one thing - when I say null value, I mean that USERID: "blah" doesn't exist on lines 1, 2, 4, and 5 in the datafile.log file. I didn't mean USERID: "". So the datafile.log actually looks like this:

TERMINAL: "blah" USERHOST: "blah"
TERMINAL: "blah" USERHOST: "blah"
TERMINAL: "blah" USERID: "kristin" USERHOST: "blah"
TERMINAL: "blah" USERHOST: "blah"
TERMINAL: "blah" USERHOST: "blah"
TERMINAL: "blah" USERID: "kristin" USERHOST: "blah"

Hope this makes sense.

I was able to find someone to help me on this. The correct code is:

perl -ne '/USERID:\s"(.+?)"/ ? print "$1\n" : print "\n"' datafile.log > userid.log