CODEWITHSUNDEEP
bashawksed
kn9
kn9

Reputation: 93

How to search and replace multi line text AWK

I have a file with following content (snippet) -- The test could be anywhere in the file.

More text here
Things-I-DO-NOT-NEED:
      name: "Orange"

      count: 8

      count: 10


Things-I-WANT:
      name: "Apple"

      count: 3

      count: 4

More text here

I would like to replace : (Including indentation) 

Things-I-WANT:
      name: "Apple"

      count: 3

      count: 4

with 

Things-I-WANT:
      name: "Banana"

      count: 7

Any suggestions on achieving it using awk/sed? Thanks!

Upvotes: 1

Views: 3149

Answers (1)

pgngp
pgngp

Reputation: 1562

You can do this in awk:

#!/usr/bin/env awk

# Helper variable
{DEFAULT = 1}

# Matches a line that begins with an alphabet
/^[[:alpha:]]+/ { 
    # This matches "Things-I-WANT:"
    if ($0 == "Things-I-WANT:") {
        flag = 1
    }
    # Matches a line that begins with an alphabet and which is 
    # right after "Things-I-WANT:" block
    else if (flag == 1) {
        print "\tname: \"Banana\""
        print ""
        print "\tcount: 7"
        print ""
        flag = 0
    }
    # Matches any other line that begins with an alphabet
    else {
        flag = 0
    }

    print $0
    DEFAULT = 0
}

# If line does not begin with an alphabet, do this
DEFAULT {
    # Print any line that's not within "Things-I-WANT:" block
    if (flag == 0) {
        print $0
    }
}

You can run this in bash using:

$ awk -f test.awk test.txt

The output of this script will be:

More text here
Things-I-DO-NOT-NEED:
      name: "Orange"

      count: 8

      count: 10


Things-I-WANT:
    name: "Banana"

    count: 7

More text here

As you can see, the Things-I-WANT: block has been replaced.

Upvotes: 3

Related Questions