Regular Expression for paired brackets

Question

The Input Line goes like this (just a part of it):

[[Text Text Text]][[text text text]]asdasdasdasda[[asdasdasdasd]]

What I want is to list all matches wherein text enclosed in a pair of [[ and ]]. I did try several patterns, but all fails when a unclosed [[ or ]] is within the input line. For example:

[[text[[text text TEXT text]]

Also, what if a single bracket exist within the input line, like:

[[text [text] text TEXT text]]

The regex pattern I used was:

\[\[[^\[\[]*\]\]

Farra

Answer 1

\[\[(?:(?!\[\[|\]\]).)*\]\]

matches [[<anything>]], where <anything> may not contain double brackets but everything else. You might need to set the DOTALL option of your regex engine to allow the dot to match newlines.

It will match [[match [this] text]] in

[[don't match this [[match [this] text]] don't match this]]

Explanation:

\[\[     # Match [[
(?:      # Match...
 (?!     #  (unless we're right at the start of
  \[\[   #   [[
 |       #  or
  \]\]   #   ]]
 )       #  )
 .       # ...any character
)*       # Repeat any number of times
\]\]     # Match ]]

Caveat: It will match [[match [this]] in the text [[match[this]]] because regexes can't count.

Steato

Answer 2

What to do with unclosed or single brackets is up to your specification. Until then

\[\[([^\]]|\][^\]])*\]\]

works for me (where the whole of the two problematic strings are matched.

Jamshedpur

@sweaver2112 You mean for a generic regex across the various implementations? The above was tested with .NET. – Jamshedpur

would there be any where escaping it would hurt? your pattern works, but on regexr it did not until i escaped the ]. it is, after all, a special char and the regex engine may not be smart enough to figure it out. – Molybdenite

Yeah, I missed the backslash for the ] after the | and I've added them for the classes too. – Jamshedpur