In a bash script I have to append a line into a systemd file. The file looks like this:
[Unit]
...
[Service]
...
The line must be appended under the [Unit] section and after all the lines in this section:
[Unit]
...
---MY LINE---
[Service]
...
I know how to append the line directly after [Unit]:
[Unit]
---MY LINE---
...
[Service]
...
with sed like this:
$ sed '/\[Unit\]/a ---MY LINE---' input_file
but how can I append my line after all the lines in the section?
If you have GNU awk in your system, you can use the following script. It appends the string right before the start of the next section.
awk -v RS='\\[[^]]*\\]' '{printf $0; if(oRT~/Unit/){print "---MY LINE---"}; printf RT; oRT=RT}' input_file
RS is the record separator. It is a regex set to anything between square bracket, such that it matches both [Unit] and [Service]. Having such a RS allows to have a record with the full content of a block section.
RT is the record terminator. It is set dynamically by awk for each record based on RS. It will hold successively [Unit] and [Service].
The trick is to store the record terminator in the variable oRT. If this one has the word Unit then append the wanted string ---MY LINE---.
If you want to append the string before the empty line at the end of the section use the following script:
awk -v RS='\\[[^]]*\\]' 'oRT~/Unit/{sub(/\n$/,"---MY LINE---\n\n")}{printf $0 RT; oRT=RT}' input_file
The sub function replaces the empty line at the end of the record by appending the wanted string and an empty line.
