[SOLVED] help with awk-print please

saeedamer · 03-16-2012, 10:01 AM

Hi there,

I need help with retrieving particular value from the output of "ps -ef" command. Here is the situation:

Code:

 ps -eaf | grep "ora_pmon_" | grep -v 'grep'

  oracle 22020342        1   0   Mar 06      -  1:11 ora_pmon_db1 
  oracle 51577060        1   0 15:49:25      -  0:05 ora_pmon_db2
  oracle 52035836        1   0   Feb 28      -  2:37 ora_pmon_db3
  oracle 66125912        1   0   Feb 28      -  1:43 ora_pmon_db4
  oracle   459068        1   0 09:52:55      -  0:00 ora_pmon_db5
  oracle  3736018        1   0   Feb 28      -  2:08 ora_pmon_db6

Out of this output, I only want the substring of values in the last column (i.e., "db1, db2, db3..."). So this is what I tried:

Code:

ps -eaf | grep "ora_pmon_" | grep -v 'grep'|awk -F' ' '{print $9}' | nawk '{print substr($1,10)}'

OUTPUT:
=======
db1 

db3
db4

db6

Please note that "db2 and db5" did not show-up in the output. Now if I change my command to the following [{print $8} instead of {print $9}], I get "db2 and db5" but all others go missing.

Code:

ps -eaf | grep "ora_pmon_" | grep -v 'grep'|awk -F' ' '{print $8}' | nawk '{print substr($1,10)}'

OUTPUT:
=======
 
db2


db5

-- end of output

I want to get all values like this:

Code:

db1
db2
db3
db4
db5
db6

Could someone please help me how this can be done?

Thank so much in advance!!

acid_kewpie · 03-16-2012, 11:00 AM

you can't use fields normally as there are a different number of whitespace elements, just cut the string at the right character using substr() instead.

danielbmartin · 03-16-2012, 11:02 AM

A brute-force approach...

Take the intermediate result you already have and do this ...

Code:

|rev          \
|cut -d_ -f1  \
|rev

... to keep whatever follows the right-most underscore.

Daniel B. Martin

firstfire · 03-16-2012, 12:26 PM

Hi.

The same as above using sed

Code:

| sed 's/.*_//'

grail · 03-16-2012, 12:28 PM

Simpler process in awk:

Code:

awk -F"_" '{print $NF}'

saeedamer · 03-16-2012, 12:40 PM

Quote:

Originally Posted by danielbmartin

A brute-force approach...

Take the intermediate result you already have and do this ...

Code:

|rev          \
|cut -d_ -f1  \
|rev

... to keep whatever follows the right-most underscore.

Daniel B. Martin

It worked - thank you much Daniel:

Code:

ps -eaf | grep "ora_pmon_" | grep -v 'grep'|rev |cut -d_ -f1 |rev

I also found this but not sure if this is as good as your suggested solution:

Code:

 ps -ef |grep 'ora_pmon_'|grep -v grep| awk '/ora_pmon/ {print $NF}'

Thanks again!

saeedamer · 03-16-2012, 12:44 PM

Thanks to Chris, grail and firstfire as well!!