I have a text file containing 100's of thousands of rows. For each line in the file, I want to remove the pipe | character if it exists between position 22 and position 57 on each line. All other pipe characters should remain in tact, just want them removed from positions 22 - 57. I know how to use sed to remove all pipes from the file but cannot figure out how to limit my search/replace to a range of characters. Can anyone help? Thanks!

One way is to cut each line into 3 sections, change only the second section, and glue the 3 lines back together ... No error checking:

#!/bin/ksh

while read line
do
len=$(expr "$line" : '.*') # length of line
f1="$(echo "$line" | cut -c1-21)"
f2="$(echo "$line" | cut -c22-57)"
f3="$(echo "$line" | cut -c58-$len)"
f2=$(echo "$f2"|sed 's/|//g')
printf "%s%s%s\n" "$f1" "$f2" "$f3"
done < data.file > newdata.file