Regex optional group parsing

Question

Been fiddling with this for hours...

I'm trying to parse error messages of this form:

[error]  C:\Me\MyPath\myFile.scala:18:22: not found: value getaa

I can do this fine with the following regex:

\[(error|warn)\]\s+(.+):(\d+):(?:\d+:)\s+(.+)$

Which correctly produces groups:

error
C:\Me\MyPath\myFile.scala
18
not found: value getaa

But to make this robust, I need to make the 22: part optional (since some versions of the scala compiler don't output column number). In other words, it needs to produce the same groups as above for this string too:

[error]  C:\Me\MyPath\myFile.scala:18: not found: value getaa

I've tried putting a question mark after the optional group, but that doesn't work - it messes up the original groups. I assume there's some stuff about lazy vs greedy that I'm not understanding. Here is a working sample on regex101. Thanks for any help.

Answer 1

You need to add two question marks:

\[(error|warn)\]\s+(.+?):(\d+):(?:\d+:)?\s+(.+)$
                      ^                ^

See a regex demo

The .+? will match any 1+ chars other than line break chars as few as possible, and will thus match up to the first occurrence of the subpatterns to follow. The second ? will make (?:\d+:) group optional.

Full pattern details

• \[ - a [
• (error|warn) - one of the two substrings (error or warn)
• \] - or just ] - a single ] char
• \s+ - 1+ whitespaces
• (.+?) - any 1+ chars other than line break chars, as few as possible, up to the first...
• : - a colon
• (\d+) - Group 2: one or more digits
• : - a colon
• (?:\d+:)? - a non-capturing group matching 1+ digits and a colon after them 1 or 0 times
• \s+ - 1+ whitespaces
• (.+) - Group 3: the rest of the line
• $ - end of string (note that it is not necessary here since .+ is a greedy subpattern)