I want to be able to retrieve a version number (n.n.n) from a file (yaml) that follows a string.

sdk: ">=0.2.4 <3.1.1"
version: ^2.3.1
sdk: flutter
version: 1.0.0+1

sed -n '/version:/p sample.yaml

which returns the line for the string provided "version:", I then piped this into another command grep to pattern match the line for the version number:

sed -n '/version:/p sample.yaml | grep -oP '(\[0-9\]+).(\[0-9\]+).(\[0-9\]+)'

I would like to see if there is a more efficient way to do this, and if string = ‘sdk’ from the example the returned value returns both when I would like to return a single value.

>Solution :

You can use

sed -nE 's/.*version:[^0-9]*([0-9][0-9.]*).*/\1/p' file
grep -oP 'version:\D*\K\d[\d.]*' file

See an online demo:

#!/bin/bash
s='sdk: ">=0.2.4 <3.1.1"
version: ^2.3.1
sdk: flutter
version 1.0.0+1'
sed -nE 's/.*version:[^0-9]*([0-9][0-9.]*).*/\1/p' <<< "$s"
# => 2.3.1
grep -oP 'version:\D*\K\d[\d.]*' <<< "$s"
# => 2.3.1

The version:\D*\K\d[\d.]* grep regex matches version: substring, then any zero or more non-digits (with \D*), then omits the matched text (with \K) and then matches and returns a digit and then any zero or more digits or dots (with \d[\d.]*).

The sed POSIX ERE (see -E option enabling this regex flavor) pattern matches

• .* – any text
• version: – a fixed substring
• [^0-9]* – zero or more non-digits
• ([0-9][0-9.]*) – Group 1 (\1): a digit and then zero or more digits or dots
• .* – any text.

The whole match is replaced with Group 1 value, the p flag prints the result of the successful substitution (-n suppresses default line output).