Unix script terminates after reading first line in file

Question

I am trying to count the number of tabs per line in file and when condition matches print the line to another file. But the script reads only the first line and terminates.

Please let me know on what's wrong in the below,

#!/bin/bash
set -e
set -o pipefail

filename="0101.tsv"
while IFS= read -r line;do

s=$(awk '{print gsub(/\t/,"")}')
echo $s

if [[ $s -eq 995 ]]; then
printf "%s\n" "$line"
continue
fi

done < $filename > abc.tsv

Thanks!.

Answer 1

You appear to want to count the number of tab-delimited fields in a file. To do this, you would have to split the input line on tabs and count them. awk can do this automatically, and it also has a special variable for the resulting number of fields, NF.

If you want to print all lines that has 996 fields (995 tabs):

awk -F '\t' 'NF == 996' <file

This is a shorthand way of writing

awk 'BEGIN { FS = "\t" } NF == 996 { print }' <file

where print means print $0, i.e. print the input record (line), and FS is the input field separator.

Whenever you find yourself extracting lines of text from a file and passing them to awk or sed or a similar tool in a loop, then there is always a more efficient way of doing the same operation. Note that the above commands only calls awk a single time, whereas your solution (had it passed the data correctly to awk) would have called awk for each and every line in the file.

Answer 2

The oneliners with awk by @steeldriver will do the job, but if you want a bash shellscript reading lines, you can do it like this,

#!/bin/bash

set -e
set -o pipefail

filename="0101.tsv"
while IFS= read -r line
do
 s=0
 len=${#line}
# echo "line=$line"
# echo "len=$len"
 for (( i=0; i<$len; i++ ))
 do
  if [ "${line:i:1}" ==  $'\t' ]
  then
   s=$((s +1))
  fi
 done
 echo $s

 if [[ "$s" == "995" ]]; then
  printf "%s\n" "$line"
  continue
 fi
done < "$filename" > abc.tsv