awk to execute external script and write the ouput to separated files 2

[ampra2]

I need help on AWK scripting to do some actions as follows:
- Parsing File,
- Do some calculation, or Execute External Program,
- Write the Output into Separated Files

I have an input file (say, input.txt) consists of this data:

834608 0.50500 openssl-sp ampra2       r     04/18/2011 09:26:18 all.q@east-32.enterprisegrid.e     1
834609 0.55500 openssl-sp ampra2       qw    04/18/2011 14:43:03                                    1
834610 0.55500 openssl-sp ampra2       qw    04/18/2011 14:43:05                                    1

I need my AWK script to read some values of it.
Those values are:
- $1 (id)
- $5 (status)
- $6 (date)
- $7 (time), and
- $8 (location)

For each record of the input file which has "qw" on $5, this script will create a file named $1.data (for each record) to store a value calculated based on the current system time and $6 + $7 on the record.

Meanwhile, for each record which has "r" as status ($5), the script will execute an external program (sh script) using those values as parameters. This external program will give a string as return value. And then this AWK script will also create a file named $1.data (for each record) to store this return value.

On this case, the AWK script will create 3 output files:
- 834608.data with return-string-value from an external program as the content.
- 834609.data and 834610.data with calculated values as the content

Anybody can help me on this problem?

Many thanks in advance.

 

[PHV]

What have YOU tried so far and where in YOUR code are you stuck ?

Hope This Helps, PH.
FAQ219-2884
FAQ181-2886

 

[feherke]

Hi
[ul]
[li]What have you tried so far ?[/li]
[li]What kind of AWK implementation can you use ?[/li]
[/ul]
The basic solution would be this :

awk '$5=="qw"{print$1,$5,$6,$7,$8>$1".data"}$5=="r"{system("echo "$1" "$5" "$6" "$7" "$8" > "$1".data")}' input.txt

But seems that your data has fixed width columns and some columns can be empty, so with [tt]gawk[/tt] you can add this to handle them :

awk '[highlight]BEGIN{FIELDWIDTHS="7 8 24 6 11 9 35 10"}[/highlight]$5=="qw"{print$1,$5,$6,$7,$8>$1".data"}$5=="r"{system("echo "$1" "$5" "$6" "$7" "$8" > "$1".data")}' input.txt

If you have no [tt]gawk[/tt] is abit lengthy, but not difficult, so I let it to you.


Feherke.

http://free.rootshell.be/~feherke/

 

[ampra2]

Hi PHV and feherke,
This is my current script.
I've made a bit change so it doesn't exactly the same thing with my description previously.

#!/bin/sh
awk '

function timedif(date,time) {
# get current system time

# calculate the difference (sec)

# convert the time difference format (hh:mm:ss)

# return the time difference

}

BEGIN {
        OFS="|";
}

{
        if ($5 == "r") {
                # run external script 
                "sendsnmp.sh $8 $1"
        } else if ($5 == "qw") {
                duration=timedif($6,$7)
                print $6, $7, $duration >> $1".qw"
}

}' input.txt

I haven't done anything for the time difference calculation, but I can write output.qw successfully if I remove some part related to the function. I'm sure if I can finish the function correctly, I can get output.qw for each record.

My problem is on the "r" one, I can't get output.run at all. Basically the script "sendsnmp.sh" will write its output to a file "output.run" where "output" is taken from the second argument (equal to $1 of input.txt). I tried executing sendsnmp.sh directly and it gives me "output.run" correctly. But when I execute this AWK script, it doesn't give me any "output.run" file. It doesn't give any error message, either.

Any idea what's happening with my script?

 

[PHV]

Replace this:
"sendsnmp.sh $8 $1"
with this:
system("sendsnmp.sh "$8" "$1)

Hope This Helps, PH.
FAQ219-2884
FAQ181-2886

 

[ampra2]

Hi PHV,

with

system("sendsnmp.sh "$8" "$1)

I get this error message:

sh: sendsnmp.sh: command not found
sh: sendsnmp.sh: command not found
sh: sendsnmp.sh: command not found

But it works when I change it to

system("./sendsnmp.sh "$8" "$1)

Thanks.

Now, I'm working on the time difference calculation.
If anybody have idea how to do this, it will help me so much.

 

[feherke]

Hi

You still not answered my question :

What kind of AWK implementation can you use ?

But as you posted no objection against the use of [tt]gawk[/tt], I assume it is fine for you :

[b]function[/b] [COLOR=darkgoldenrod]timediff[/color][teal]([/teal]date[teal],[/teal]time   [teal],[/teal]diff[teal])[/teal]
[teal]{[/teal]
  diff[teal]=[/teal][COLOR=chocolate]systime[/color][teal]()-[/teal][COLOR=chocolate]mktime[/color][teal]([/teal][COLOR=chocolate]gensub[/color][teal](/([[:[/teal]digit[teal]:]]+)\/([[:[/teal]digit[teal]:]]+)\/([[:[/teal]digit[teal]:]]+)/,[/teal][green][i]"[/i][/green][lime][i]\\[/i][/lime][green][i]3 [/i][/green][lime][i]\\[/i][/lime][green][i]1 [/i][/green][lime][i]\\[/i][/lime][green][i]2"[/i][/green][teal],[/teal][green][i]""[/i][/green][teal],[/teal]date[teal])[/teal] [green][i]" "[/i][/green] [COLOR=chocolate]gensub[/color][teal](/:/,[/teal][green][i]" "[/i][/green][teal],[/teal][green][i]"g"[/i][/green][teal],[/teal]time[teal]))[/teal]
  [b]return[/b] [COLOR=chocolate]int[/color][teal]([/teal]diff[teal]/[/teal][purple]24[/purple][teal]/[/teal][purple]60[/purple][teal]/[/teal][purple]60[/purple][teal])[/teal] [green][i]" days "[/i][/green] [COLOR=chocolate]strftime[/color][teal]([/teal][green][i]"%T"[/i][/green][teal],[/teal]diff[teal],[/teal][purple]1[/purple][teal])[/teal]
[teal]}[/teal]

Feherke.

http://free.rootshell.be/~feherke/