Replace string in specific line at specific position of fixed-length file

Question

I have a fixed-length data file where records are tagged (first 6 characters have a certain string that identifies the record) and I need to replace text from a specific line and a specific position on that line.

HEADER   123456
SHIPTO   CODE    123 LANE     HOUSTON    TX
ITEM     ACME BRICK

For example, I would like to replace the ship-to code in the above data. The data can be found in positions 10-17 of the ship-to record (line starts with "SHIPTO"). I'd like to use sed or awk (or other commands) but update the data with a single command line. If there is more than one SHIPTO line, I am ok with updating them all.

Answer 1

Using sed:

$ sed -E '/^SHIPTO/ s/(.{9}).{8}/\1NuValue /' file
HEADER   123456
SHIPTO   NuValue 123 LANE     HOUSTON    TX
ITEM     ACME BRICK

How it works:

• /^SHIPTO/

  This selects lines that start with SHIPTO. The substitute command which follows will only be applied to these lines.

• s/(.{9}).{8}/\1NuValue /

  This matches the first 9+8 characters of the line and saves the first 9 characters of the line in group 1. These characters are replaced by group 1, \1, and your new value (which, to keep the format, should be 8 characters.

Answer 2

Given the input file:

HEADER   123456
SHIPTO   CODE    123 LANE     HOUSTON    TX
ITEM     ACME BRICK

And the desire to change the field after 'CODE' where the first field is 'SHIPTO':

$ awk 'BEGIN { OFS = "\t" } $1 != "SHIPTO" { print } $1 == "SHIPTO" { $3="0000"; print }' input
HEADER   123456
SHIPTO  CODE    0000    LANE    HOUSTON TX
ITEM     ACME BRICK

If your fields are fixed-length and space-delimited and not tab delimited you can use formatted print strings (e. g. printf( "%20s\t" "$1" )).