In a shell script, how do I replace the first occurrence of a string, after a different string

Question

I have a simple config file that looks a bit like this:

[sectionA]
url = value1
username = value2
password = value3

[sectionC]
url = value1
username = value2
password = value3

[sectionB]
url = value1
username = value2
password = value3

And I want to replace the username for SectionB to be valueX without touching SectionA's or SectionC's username.

I've tried some variations on sed, but as far as I've managed to fathom it seems to operate on individual lines.

How do I do the equivalent of

Search for StringA (in this case [SectionB])
Find the next occurrence of StringB (username = value2)
Replace with StringC ('username = valueX`)

This is possible using the shell, but would be much easier using a language with an ini parsing library. — jordanm, Aug 15 '12 at 23:06

score 4 · Accepted Answer · answered Aug 15 '12 at 23:26

4

`sed`:

sed '/sectionB/,/\[/s/username.*/username = valueX/' input

`awk`:

awk -vRS='' -vFS='\n' -vOFS='\n' '
$1~/sectionB/{
    sub(/=.*$/, "= valueX", $3)
}
{
    printf "%s\n\n", $0
}' input

answered Aug 15 '12 at 23:26

kev

155,172
47
273
272

I had no idea it could be so simple. Nice +1 – mVChr Aug 15 '12 at 23:32
1

Perfect, thanks. Some of the sed syntax is a little... cryptic. – Cylindric Aug 16 '12 at 19:23

score 0 · Answer 2 · answered Aug 15 '12 at 23:28

0

This multi-line sed should do the trick:

sed -E -n '1h;1!H;${;g;s/\[sectionB\]([^[]*)username = [a-zA-Z0-9_]+/\[sectionB\]\1username = valueX/g;p;}' input.txt

answered Aug 15 '12 at 23:28

mVChr

49,587
11
107
104

In a shell script, how do I replace the first occurrence of a string, after a different string

2 Answers2

sed:

awk:

`sed`:

`awk`: