naming output file in while loop

Question

I have a file input.txt with three columns

A B Z
C D Y
E F X

I want to read the first two fields, perform some function and name the output with the last field.

while IFS=" " read -r f1 f2 f3; do cat $f1 $f2 > $f3; done < file.txt

How can I apply this to longer codes

the input.sample.text has following three columns

../Data/G20P1sc-C05-R1.fastq.gz ../Data/G20P1sc-C05-R1.fastq.gz G20P1sc-C05
../Data/G20P1sc-C08-R1.fastq.gz ../Data/G20P1sc-C08-R1.fastq.gz G20P1sc-C08

 while IFS=" " read -r f1 f2 f3; do rsem-calculate-expression --star --star-path /share/pkg/star/2.7.0e/bin --temporary-folder --star-gzipped-read-file --paired-end $f1 $f2 ../genome_indices/rsem-star/rsem-star ../test_results/$f3; done < src_pdm3012/input.sample.text 

The output has the following error

EXITING: FATAL INPUT ERROR: empty value for parameter "outFileNamePrefix" in input "Command-Line-Initial"
SOLUTION: use non-empty value for this parameter

I try to explicitly use the --outFileNamePrefix and it still has the same error

while IFS=" " read -r f1 f2 f3; do rsem-calculate-expression --star --star-path /share/pkg/star/2.7.0e/bin --temporary-folder --star-gzipped-read-file --paired-end $f1 $f2 ../genome_indices/rsem-star/rsem-star --outFileNamePrefix $f3; done < ../src_pdm3012/input.sample.text

Answer 1

I suggest awk script to do this:

awk '{ print $1,$2 > $3}'

Of course you must be sure that $3 exists (and is unique) and that there is no confution what the third column is, i.e. columns 1, 2, and 3 should not contain field separators.

awk will also allow you manyfold manipulations on the first two columns.

Answer 2

You can redirect a command group or any form of compound command to redirect the output of more than one command, whether it's a while loop or not:

while IFS=" " read -r f1 f2 f3; do
  {
    cat < "$f1"
    echo whatever
    cat < "$f2"
    echo end
  } > "$f3"
done < file.txt

Also remember that in POSIX shells, variables expansions need to be quoted.