[UnIxBug]

Hi all,

waht I am trying to do is to invoke an awk inside another, using the system() call from awk (along with other commands) but it will return only errors like "^ unterminated string" , "^ unexpected newline", or syntaxes errors.

The small script I am trying to do would look like:

---------------

#!/bin/bash
#!/bin/awk -f


ls -lrt /var/log * >& 1 | egrep "log" | sort -m | awk '{print strftime("%y %m %d")"\t", "/var/log/"$9"\t",
substr($8,3,4)"\t", system("cal -y | awk 'NF==3 {printf "%s\n%s\n%s\n",$1,$2,$3}' | cat -n | grep $6 | cut 1")"\t", $7}'

----------------


But if executed returns the following error:

awk: cmd. line:2: substr($8,3,4)"\t", system("cal -y | awk NF==3
awk: cmd. line:2: ^ unterminated string



Both sentences, if executed apart, work:

$ls -lrt /var/log * >& 1 | egrep "log" | sort -m | gawk '{print strftime("%y %m %d")"\t", "/var/log/"$9"\t", substr($8,3,4)"\t", $6"\t", $7}'

would give as output:


...
09 02 24 /var/log/[logfilename] 06 Mar 25
...


and

$cal -y | awk 'NF==3 {printf "%s\n%s\n%s\n",$1,$2,$3}' | cat -n | grep Feb | cut -f1

transforms the abbreviated month into his numeric

$2

What I am intending to do is that for each output of the first awk, the abbreviated month that the "ls -lrt" returns is transformed into its numerical by the second command sentence:


09 02 24 /var/log/[logfilename] 06 Mar 25 -----------> 09 02 24 /var/log/[logfilename] 06 3 25


Any idea why the second awk does not work inside trhe system() call or anyone has a suggestion on achieving the conversion of the month from abbreviated to numeric?

Thanks in advance for any help!

[eggi]

Hey there,

Actually, if i'm not mistaken, even if you get that to work, awk's system function only sends back the return code (did it work or not), won't help you.

One thing you could do is to assign the output of your nested awk command to a shell variable and then use awk's -v flag to pass that variable to your outer awk script - something like this (can't test it, but it should give you the idea

variable=`awk $cal -y | awk 'NF==3 {printf "%s\n%s\n%s\n",$1,$2,$3}' | cat -n | grep Feb | cut -f1 SOME_FILE_NAME`

awk -v var=variable '{print strftime("%y %m %d")"\t", "/var/log/"$9"\t", var, substr($8,3,4)"\t", $6"\t", $7}'

It's not even close to working the way you want, just an example of how can insert awk into awk

Best wishes,

Mike

[UnIxBug]

In the end puting the field with month between simple quotes aka '$6' transforms it into numerical

ls -lrt /var/log * >& 1 | egrep "log" | sort -m | awk '{print strftime("%y %m %d")"\t", "/var/log/"$9"\t",
substr($8,3,4)"\t", '$6' "\t", $7}'

[eggi]

nteresting - not what I would have expected - cool find

, Mike

UnIxBug, on Feb 25 2009, 06:04 AM, said:

In the end puting the field with month between simple quotes aka '$6' transforms it into numerical

ls -lrt /var/log * >& 1 | egrep "log" | sort -m | awk '{print strftime("%y %m %d")"\t", "/var/log/"$9"\t",
substr($8,3,4)"\t", '$6' "\t", $7}'