BASH

carlolab · 04-11-2020, 07:50 AM

Hi,

I have a "fileX" test file with a content like this:

a b c d e f

I need to make a for cicle to print column 2 to 6

for col in {2..6}
do
cat fileX | awk '{ print $$col }'
done

The problem is to pass variable $col as variable in print function of awk.
The part '{ print $$col }' shold be read as '{ print $2 }' then '{ print $3 }' etc.etc.

Can someone help me?

Thank You

Carlo

Turbocapitalist · 04-11-2020, 08:14 AM

You could do the loop within AWK, or you could take advantage of the -v option to pass an environment variable into the AWK script.

Code:

#!/bin/sh

for col in {2..6}
do
        cat fileX | awk -v c=col '{ print $c }'
done

exit 0

That will print the different columns, from 2 through 6, on different lines.

If you want them all on the same line then put the loop inside AWK.

Code:

cat fileX \
| awk '{ for (col=2;col<=6;col++) { printf("%s%s",$col,OFS);} print "";}'

See "man awk" for either method. It is a short manual.

carlolab · 04-11-2020, 08:26 AM

Quote:

Originally Posted by Turbocapitalist

You could do the loop within AWK, or you could take advantage of the -v option to pass an environment variable into the AWK script.

Code:

#!/bin/sh

for col in {2..6}
do
        cat fileX | awk -v c=col '{ print $c }'
done

exit 0

That will print the different columns, from 2 through 6, on different lines.

If you want them all on the same line then put the loop inside AWK.

Code:

cat fileX \
| awk '{ for (col=2;col<=6;col++) { printf("%s%s",$col,OFS);} print "";}'

See "man awk" for either method. It is a short manual.

Thank you for answering.

I tried but it didn't work.

#!/bin/sh

for col in {2..6}
do
cat fileX | awk -v c=col '{ print $c }'
done

exit 0

It print:

a b c d e f
a b c d e f
a b c d e f
a b c d e f
a b c d e f

Carlo

Turbocapitalist · 04-11-2020, 08:31 AM

My mistake, I missed that {2..6} is a bashism. Change the first line, and add in the dollar sign I forgot while we're in there.

Code:

#!/bin/bash

for col in {2..6}
do
        cat fileX | awk -v c=$col '{ print $c }'
done

exit 0

and it will produce a different result. I typed #!/bin/sh by reflex.

carlolab · 04-11-2020, 08:38 AM

Quote:

Originally Posted by Turbocapitalist

My mistake, I missed that {2..6} is a bashism. Change the first line, and add in the dollar sign I forgot while we're in there.

Code:

#!/bin/bash

for col in {2..6}
do
        cat fileX | awk -v c=$col '{ print $c }'
done

exit 0

and it will produce a different result. I typed #!/bin/sh by reflex.

It works!

Thank You very much Turbocapitalist!

Carlo

pan64 · 04-11-2020, 09:50 AM

you can read the file as an array in bash:

Code:

read -r -a arr <fileX
printf "%s\n" "${arr[@]}"

(not tested)

from the other hand you do not need cat, you can simply use:

Code:

awk 'script' fileX

rnturn · 04-13-2020, 10:05 PM

Quote:

Originally Posted by carlolab

Code:

for col in {2..6}
do
cat  fileX | awk '{ print $$col }'
done

The problem is to pass variable $col as variable in print function of awk.
The part '{ print $$col }' shold be read as '{ print $2 }' then '{ print $3 }' etc.etc.

My first thought was that you should use braces: "$$col" -> "${${col}}" or "${$col}" to avoid something like:

Code:

$ echo $$col
12345col

But I think your biggest problem would be substituting that into your awk command because of the single quotes and getting the different values of `col' that you need.

Perhaps building the awk command string up a bit at a time and issuing "eval awk $CMDSTR" would work.

Or doing the whole thing in awk (sorry... been too long since I wrote much awk code).

NevemTeve · 04-13-2020, 10:57 PM

@OP: The actual input-file contains only one line? If not, then please provide a more complete example.

shruggy · 04-14-2020, 02:52 AM

Just expanding a bit on pan64's solution from #6. You can slice arrays in bash, too:

Code:

printf %s\\n "${arr[@]:1}"