Re: bash for loop and awk

On Fri, Dec 23, 2011 at 2:25 PM, Marshall Davis <[email protected]> wrote:
> Hey all,
> I'm having some trouble figuring out how to escape variables inside a bash
> for loop used as field names in awk.
> Here is the basic idea: get the max value from each column in a csv file
> with a number of columns less than 100. This is what I have tried so far:
>
> sed -i 's/,/ /g' ./file.csv
> for fieldnum in {1..100}
>     do
>     awk '{if(min==""){min=max=$$fieldnum}; if($$fieldnum>max)
> {max=$$fieldnum};} END {print max}' ./file.csv
>     done
>

I don't think this is going to work like you think it should.

Putting the $fieldnum inside single quotes in the awk string prevents
it from being expanded by the shell.
You can see this by putting:   echo $fieldnum; echo '$fieldnum' and
you will see two different results.

Try this:

for fieldnum in {1..100}; do
awk -v field="$fieldnum"
'{if(max==""){max=$field};if($field>max){max=$field};} END {print
max}' ./file.csv
done

the "-v" argument to awk does the bash to awk variable substitution.

I hope this helps,

--Donald


> I have tried escaping the $ and $fieldnum with backslash, parentheses,
> quotes, I can't use single quotes, as that is what awk uses for it's
> expressions, and every way I have tried that doesn't give a syntax error
> just prints the entire line $fieldnum times. Is there a way to do this, or
> should I look for a tool other than awk? I wouldn't be opposed to using a
> different shell, although I only have experience with bash and a smidgen
> with korn.
> Thanks!
> Marshall



-- 
Donald Cowart
http://www.rdex.net/

---------------------------------------------------------------------
Archive      http://marc.info/?l=jaxlug-list&r=1&w=2
RSS Feed     http://www.mail-archive.com/[email protected]/maillist.xml
Unsubscribe  [email protected]