Hey guys, Looking through my book and the web and I am not having any success returning data from a search.
I need to have awk search for a string and print the first field which is no problem but now its returning two options as the input data has changed. The change is reliable, I only want the first field if it ends in a regex that I have, and I only want what that regex matches to be printed. Is it possible to do this in a one liner so I don't need to construct an awk script?
I suppose I could pipe it into grep and cut but that's not very sexy :)
Any ideas?
Thanks! jlc
On Wed, Jun 24, 2009 at 10:31:21PM +0000, Joseph L. Casale wrote:
I need to have awk search for a string and print the first field which is no problem but now its returning two options as the input data has changed. The change is reliable, I only want the first field if it ends in a regex that I have, and I only want what that regex matches to be printed. Is it possible to do this in a one liner so I don't need to construct an awk script?
What you've written is mostly incoherent and incomprehensible.
Free clue: provide examples of input and expected output, including examples that will _not_ match (so no false positives occur).
I'm not sure if you simply want awk '$1 ~ /regexp/ { print $1}' or something more.
What you've written is mostly incoherent and incomprehensible.
heh, been a long day:)
Actually I am using gawk "/string1/ && /string2/ { print substr( $1, length($1) - 1, length($1) ) }" which gets me what I need, the last two characters of the first field of a specific match...
Sorry <vbg>, jlc
On Wed, Jun 24, 2009 at 11:15:10PM +0000, Joseph L. Casale wrote:
What you've written is mostly incoherent and incomprehensible.
heh, been a long day:)
Actually I am using gawk "/string1/ && /string2/ { print substr( $1, length($1) - 1, length($1) ) }" which gets me what I need, the last two characters of the first
The final parameter to substr() isn't needed (and would only be "2" for what you wanted)
field of a specific match...
Well, it's only printing the first field, not necessarily the field that matched. eg $ echo hello there everyone | awk '/there/ && /everyone/ { print substr($1,length($1)-1) }' lo
You need to do $1 ~ /regex/ if you only want to test against the first field
Joseph L. Casale wrote:
Hey guys, Looking through my book and the web and I am not having any success returning data from a search.
I need to have awk search for a string and print the first field which is no problem but now its returning two options as the input data has changed. The change is reliable, I only want the first field if it ends in a regex that I have, and I only want what that regex matches to be printed. Is it possible to do this in a one liner so I don't need to construct an awk script?
I suppose I could pipe it into grep and cut but that's not very sexy :)
Sed is the next tool up from grep:
sed -n -e 's/.*(regex.*)/\1/p' will print the matching regex, so if you can expand it to match whatever you are calling a field it should work.
Joseph L. Casale <JCasale@...> writes:
Hey guys, Looking through my book and the web and I am not having any success returning data from a search.
I need to have awk search for a string and print the first field which is no problem but now its returning two options as the input data has changed. The change is reliable, I only want the first field if it ends in a regex that I have, and I only want what that regex matches to be printed. Is it possible to do this in a one liner so I don't need to construct an awk script?
I suppose I could pipe it into grep and cut but that's not very sexy :)
Any ideas?
Thanks! jlc
Being an old perl hacker I have to at least suggest doing whatever you're attempting in perl. perl gives you much more powerful and flexible regular expression processing. It also makes it really simple to pull out whatever matched within the RE.
Cheers, Dave
Being an old perl hacker I have to at least suggest doing whatever you're attempting in perl.
Yeah I agree but I have found myself needing to duplicate so much work moving between nix and windows that I resolve to just using UnxTools under windows so I stay with shell scripting so I can move work back and forth. I don't always have the luxury of installing Perl in windows.
With the portable vanilla Perl distro coming along that may change! I do know a smidge of Perl and I love it, its easy to do some great things with.
jlc
David G. Miller wrote:
Hey guys, Looking through my book and the web and I am not having any success returning data from a search.
I need to have awk search for a string and print the first field which is no problem but now its returning two options as the input data has changed. The change is reliable, I only want the first field if it ends in a regex that I have, and I only want what that regex matches to be printed. Is it possible to do this in a one liner so I don't need to construct an awk script?
I suppose I could pipe it into grep and cut but that's not very sexy :)
Any ideas?
Thanks! jlc
Being an old perl hacker I have to at least suggest doing whatever you're attempting in perl. perl gives you much more powerful and flexible regular expression processing. It also makes it really simple to pull out whatever matched within the RE.
This case is probably simple enough for sed, but in general I agree that if you need awk you probably might as well use perl which can also probably do a better job than the rest of the shell script that is likely surrounding this operation.
-
David G. Miller -
Joseph L. Casale -
Les Mikesell -
Stephen Harris