Can awk be used to call shell programs (cat, cp, rm, etc.)

kmkocot · 11-16-2009, 10:25 AM

Hi all,

I am trying to write a simple awk script to remove a file (infile) if another file containing information about it (infile.count) contains a value >5 anywhere in $2.

Code:

awk -F" " '{if ($2>5) rm infile}' infile.count

When I run the script as is, nothing happens. When I change rm infile to print $2 it behaves normally. Is there a way to get awk to call rm or another shell program? Otherwise, can anyone give me me pointers in writing a script that removes infile if infile.count contains any integers greater than 5?

infile.count looks like this:

Code:

ACAL 6
APOM 29
OCT1 3
MGAL 2
CNIT 1
CAPI 1
NVEC 1
HROB 5

Thanks!
Kevin

catkin · 11-16-2009, 10:46 AM

You can use awk's built-in function "system()". Here's an example from The GNU Awk User's Guide

For example, if the following fragment of code is put in your awk program:

Code:

END {
    system("date | mail -s 'awk run done' root")
}

the system administrator is sent mail when the awk program finishes processing input and begins its end-of-input processing.

kmkocot · 11-17-2009, 09:05 PM

Thanks for the help but I'm still having a problem. For these examples I renamed my infile that I wanted to delete to lamb.txt to demonstrate a problem I'm having.

With quotes around the filename I get this (the file does exist):

Code:

awk -F " " '{if ($2>5) {system(rm "lamb.txt")}}' orthology_4_test.txt
sh: lamb.txt: not found

Without quotes around the filename I get this:

Code:

awk -F " " '{if ($2>5) {system(rm lamb.txt)}}' orthology_4_test.txt
awk: {if ($2>5) {system(rm lamb.txt)}}
awk:                           ^ syntax error
awk: fatal: 0 is invalid as number of arguments for system

Can you tell me what I'm doing wrong now? I can't find many examples using system but I assume rm is a legal program to call using it?

Thanks!
Kevin

ghostdog74 · 11-17-2009, 09:10 PM

the error says it, do you actually have lamb.txt in the current path?

kmkocot · 11-17-2009, 11:53 PM

Absolutely. I really don't understand what the problem is.

ghostdog74 · 11-18-2009, 12:14 AM

you have two $2 values that are more than 5. When the 2nd record that has $2 > 5 is processed, the lamb.txt is already deleted. that's why it cannot find the file!

kmkocot · 11-19-2009, 12:37 AM

I neglected to mention the fact that the file isn't actually being deleted.

Vrajgh · 11-19-2009, 04:06 AM

Try:

Code:

$ awk -F " " '{if ($2>5) {system("rm lamb.txt")}}'

catkin · 11-19-2009, 05:16 AM

Quote:

Originally Posted by kmkocot

With quotes around the filename I get this (the file does exist):

Code:

awk -F " " '{if ($2>5) {system(rm "lamb.txt")}}' orthology_4_test.txt
sh: lamb.txt: not found

In this case the string passed to the system function is the awk expression rm "lamb.txt". The variable rm is not set so the expression is evaluated to lamb.txt and the result is as if you had entered that at the prompt.

Quote:

Originally Posted by kmkocot

Without quotes around the filename I get this:

Code:

awk -F " " '{if ($2>5) {system(rm lamb.txt)}}' orthology_4_test.txt
awk: {if ($2>5) {system(rm lamb.txt)}}
awk:                           ^ syntax error
awk: fatal: 0 is invalid as number of arguments for system

In this case the string passed to the system function is the awk expression rm lamb.txt. The . in lamb.txt is not syntactically valid.

Vrajgh's suggestion should work.

david1941 · 11-19-2009, 06:05 AM

You are actually setting up the command line in awk and THEN executing it. Try:

Code:

{print"/bin/rm","lamb.txt"; system("")}