regexp: multiline, non-greedy match until optional string

Question

Using Go's regexp, I'm trying to extract a predefined set of ordered key-value (multiline) pairs whose last element may be optional from a raw text, e.g.,

 Key1:
  SomeValue1
  MoreValue1
 Key2:
  SomeValue2
  MoreValue2
 OptionalKey3:
  SomeValue3
  MoreValue3

(here, I want to extract all the values as named groups)

If I use the default greedy pattern (?s:Key1:\n(?P<Key1>.*)Key2:\n(?P<Key2>.*)(?:OptionalKey3:\n(?P<OptionalKey3>.*))?), it never sees OptionalKey3 and matches the rest of the text as Key2.

If I use the non-greedy pattern (?s:Key1:\n(?P<Key1>.*)Key2:\n(?P<Key2>.*?)(?:OptionalKey3:\n(?P<OptionalKey3>.*))?), it doesn't even see SomeValue2 and stops immediately: https://regex101.com/r/QE2g3o/1

Is there a way to optionally match OptionalKey3 while also able to capture all the other ones?

score 2 · Accepted Answer · answered Jun 23 '21 at 22:34

Use

(?s)\AKey1:\n(?P<Key1>.*)Key2:\n(?P<Key2>.*?)(?:OptionalKey3:\n(?P<OptionalKey3>.*))?\z

See regex proof.

EXPLANATION

--------------------------------------------------------------------------------
  (?s)                     set flags for this block (with . matching
                           \n) (case-sensitive) (with ^ and $
                           matching normally) (matching whitespace
                           and # normally)
--------------------------------------------------------------------------------
  \A                       the beginning of the string
--------------------------------------------------------------------------------
  Key1:                    'Key1:'
--------------------------------------------------------------------------------
  \n                       '\n' (newline)
--------------------------------------------------------------------------------
  (?P<Key1>                 group and capture to "Key1":
--------------------------------------------------------------------------------
    .*                       any character (0 or more times (matching
                             the most amount possible))
--------------------------------------------------------------------------------
  )                        end of "Key1"
--------------------------------------------------------------------------------
  Key2:                    'Key2:'
--------------------------------------------------------------------------------
  \n                       '\n' (newline)
--------------------------------------------------------------------------------
  (?P<Key2>                group and capture to "Key2":
--------------------------------------------------------------------------------
    .*?                      any character (0 or more times (matching
                             the least amount possible))
--------------------------------------------------------------------------------
  )                        end of "Key2"
--------------------------------------------------------------------------------
  (?:                      group, but do not capture (optional
                           (matching the most amount possible)):
--------------------------------------------------------------------------------
    OptionalKey3:            'OptionalKey3:'
--------------------------------------------------------------------------------
    \n                       '\n' (newline)
--------------------------------------------------------------------------------
    (?P<OptionalKey3>         group and capture to "OptionalKey3":
--------------------------------------------------------------------------------
      .*                       any character (0 or more times
                               (matching the most amount possible))
--------------------------------------------------------------------------------
    )                        end of "OptionalKey3"
--------------------------------------------------------------------------------
  )?                       end of grouping
--------------------------------------------------------------------------------
  \z                       the end of the string

Ah, the trick is to add `\z` at the end so that it forces the entire text to be matched. Thanks! — Joseph, Jun 23 '21 at 22:44

regexp: multiline, non-greedy match until optional string

1 Answers1