[SOLVED] Sed command: how to search for the second line occurrence starting from a word?

Issue

I have this file simple.yml:

services:
  - name: first
    image: first_image
    versionId: 1.0.0_first_image_id
  - name: second
    image: second_image
    versionId: 2.0.1_second_image_id

I'm trying to replace the second occurrence of versionId with this value: versionId: image_id_edited so that the file would look like this:

services:
  - name: first
    image: first_image
    versionId: 1.0.0_first_image_id
  - name: second
    image: second_image
    versionId: image_id_edited

The value 2.0.1_second_image_id here is changed often, so I cannot do a static search. Currently, I have this sed command but it's not working:

sed -rz "s/versionId:\s*.*\$/versionId: image_id_edited/2" -i simple.yaml

I'd like to have a regex to search for the second versionId: ... until it reaches the end of the line! Thanks!

Solution

You appear to be counting on the -z flag to cause the whole input to be interpreted as a single line, so that the 2 flag to your s command results in the second match in the whole file being substituted. The problem is that the relevant count is over non-overlapping matches. Your pattern being anchored to the end of the line, you need to match in multiline mode in order for there to be a possibility of two non-overlapping matches:

sed -rz 's/versionId:\s*.*$/versionId: image_id_edited/m2' -i simple.yaml

That is, add an m flag to the s command.

Answered By - John Bollinger

Answer Checked By - Mary Flores (WPSolving Volunteer)