Question

I have a small problem and I would appreciate helping me in it.

In summary, I have a file:

1,5,6,7,8,9

2,3,8,5,35,3

2,46,76,98,9

I need to read specific lines from it and print them into another text document. I know I can use (awk '{print "$2" "$3"}') to print the second and third columns beside each other. However, I need to use two statement as (awk '{print "$2"}' >> file.text) then (awk '{print "$3"}' >> file.text), but the two columns would appear under each other and not beside each other.

How can I make them appear beside each other?

Answer 1

If you must extract the columns in separate processes, use paste to stitch them together. I assume your shell is bash/zsh/ksh, and I assume the blank lines in your sample input should not be there.

paste -d, <(awk -F, '{print $2}' file) <(awk -F, '{print $3}' file)

produces

5,6
3,8
46,76

Without the process substitutions:

awk -F, '{print $2}' file > tmp1
awk -F, '{print $3}' file > tmp2
paste -d, tmp1 tmp2 > output

---

Update based on your answer:

On first appearance, that’s a confusing setup. Does this work?

for (( x=1; x<=$number_of_features; x++ )); do
    feature_number=$(sed -n "$x {p;q}" feature.txt)
    if [[ -f out.txt ]]; then
        cut -d, -f$feature_number file.txt > out.txt
    else
        paste -d, out.txt <(cut -d, -f$feature_number file.txt) > tmp &&
        mv tmp out.txt
    fi
done

That has to read the file.txt file a number of times. It would clearly be more efficient to only have to read it once:

awk -F, -f numfeat=$number_of_features '
    # read the feature file into an array
    NR==FNR {
        colno[++i] = $0
        next
    }

    # now, process the file.txt and emit the desired columns
    {
        sep = ""
        for (i=1; i<=numfeat; i++) {
            printf "%s%s", sep, $(colno[i])
            sep = FS
        }
        print ""
    }
' feature.txt file.txt > out.txt