UNIX for Dummies Questions & Answers

Hey guys, I'm sure someone will have asked this before but I searched for ages and couldn't find a post.

Anyhow, I have a datafile and I want to search for "UserID : " in it and then return the characters following this until it hits a pipe.

e.g.

Running the comand against the following data:

[01/01/2008] [INFO] |BLAH BLAH|UserID : 12345| RANDOMTEXT
[01/01/2008] |BLAH BLAH|UserID : 54321| RANDOMTEXT | RANDOMTEXT
[INFO] |BLAH BLAH|UserID : 666| RANDOMTEXT

Would return:

12345
54321
666

If any of you have an idea how to do this I would be most grateful. I'm using a Solaris box by the way. Thanks

Code:

cut -d": " -f 2 filename | cut -d"|" -f 1

Hmmmm, thanks but using cut -d": " -f 2 filename | cut -d"|" -f 1

... I get "cut: invalid delimiter"

I tried it without the space after the ":" but unfortunately some of the data includes ":" that aren't in "UserID : " (admittedly I didn't put that in my example data so I'll change it now)

[01:01:2008] [INFO] |BLAH BLAH|UserID : 12345| RANDOMTEXT
[01:01:2008] |BLAH BLAH|UserID : 54321| RANDOMTEXT | RANDOMTEXT
[INFO] |BLAH BLAH|UserID : 666| RANDOMTEXT

Any thoughts?

I should do something like this with sed:

Code:

sed 's/.*: \(.*\)|.*/\1/' file

Regards

Code:

sed -n 's/.*|UserID : \([^|]*\)|.*/\1/p'

Worked a treat. Thanks for you help (again!)

As an awk fan I thought I'd offer this little snippet which pulls out the whole field for manipulation rather than simply the string following the 'userid:'

>cat rant
sdf|asdf|asfgadg|shddh|userid:sdg|sdgdg|sfhdfs|
sdf|asdf|asfgadg|shddh|userid:sdg|sdgdg|sfhdfs|
sdf|asdf|asfgadg|shddh|userid:sdg|sdgdg|sfhdfs|
...
>awk 'BEGIN {RS="|"; FS=":"} /userid/ {print $2}' rant
sdg
sdg
sdg
...