topic Re: script help in Operating System - HP-UX

script help

Jeff Paciolla_1 — Wed, 10 Oct 2001 16:49:20 GMT

Does anybody have any example of how to parse text by pattern using awk, sed, or grep. I have tried every combination of script that I can think of.

Here is my latest attempt:
test="abcd aa abc : rylfrer01 (xxx.xx.x.xxx) unreachable."

In my script i am trying the following

data=`echo "$test"|sed 's/.*(\: .*\ ).*/\1/'`
echo "$data"

I was hoping this would give me the text rylfrer01

Thanks for your help.
Jeff

Re: script help

Patrick Wallek — Wed, 10 Oct 2001 16:59:55 GMT

What is it you're trying accomplish?

Re: script help

Sridhar Bhaskarla — Wed, 10 Oct 2001 17:03:18 GMT

If the example given is the format of your input, then it's easy. Did you try this?

date=`echo $test|awk'{FS=":";print $2}' |awk '{print $1}'`

This will give you rfylfer01.

-Sri

Re: script help

Sridhar Bhaskarla — Wed, 10 Oct 2001 17:05:51 GMT

Ooops... it's not date..it's data. Replace your sed statement with the statement I gave
by replacing date with data.

-Sri

Re: script help

James R. Ferguson — Wed, 10 Oct 2001 17:07:49 GMT

Hi Jeff:

# echo $test|awk '{print $5}'

Regards!

...JRF...

Re: script help

Jeff Paciolla_1 — Wed, 10 Oct 2001 17:08:47 GMT

Patrick - What I am trying to get is the text that starts with "ryl*" from the test message. My dilemma is that the pattern text "ryl*" does not always appear in the same location within the message. So in my script I am trying to parse the ryl* text no matter when or where it appears in the message, and assign that to a different variable.

Thanks
Jeff

Re: script help

A. Clay Stephenson — Wed, 10 Oct 2001 17:31:52 GMT

Okay Jeff, here's my cut at it in awk but it's not real pretty but it will find the ryl string:

TEST="abcd aa abc: rylfrer01 (xxx.xx.x.xxxx) unreachable"

echo ${TEST} | awk '{n = split($0,arry," "); if (n > 0) { i = 1; while (i <= n) {if (match(arry[i],/ryl/) > 0) {printf("%s\n",arry[i])}; ++i }}}'

Clay

Re: script help

Wodisch — Wed, 10 Oct 2001 17:44:38 GMT

Hello,

how about plain ol' "cut"?

echo $Test | cut -d':' -f2 | cut -d'(' -f1

Just my ?0.02,
Wodisch

Re: script help

Jeff Paciolla_1 — Wed, 10 Oct 2001 17:55:01 GMT

Thanks Clay - This seems to work perfectly

Jeff

Re: script help

Sridhar Bhaskarla — Wed, 10 Oct 2001 18:05:50 GMT

Hopefully, I understood your question now.

Try this.

data=`echo $test|sed 's/^.*ryl/ryl/'` |awk "{print $1}'

if you want the entire text after ryl.., take out the awk statement.

-Sri

Re: script help

Curtis Larson_1 — Wed, 10 Oct 2001 19:28:10 GMT

echo "$test" | sed 's/.*$rly.*$[[:space:]].*/\1/'