正则表达式在相同的输入上返回不同的结果

Question

I am checking for some results and passing two inputs from a document, two inputs looks same but why are they returning different output. my regex is

(?<preandconjunct>(?:\b([Ss]ubsection|[Ss]ection|[Aa]rticle) +)(?<conjunct>(?:(?<level>(?:(?:[IVXivx]{1,5}(?![A-Z]))|(?:[A-Z]{1,2}(?![A-Z]))|(?:[0-9]+)))|(?<level>\((?:(?:[IVXivx]{1,5}(?![A-Z]))|(?:[A-Z]{1,2}(?![A-Z]))|(?:(?!in|or|if|of|to|as|at|it|no|an)[a-z]{1,2}(?![a-z]))|(?:[0-9]+))\))|(?<level>[\.-](?:(?:[IVXivx]{1,5}(?![A-Z]))|(?:[A-Z]{1,2}(?![A-Z]))|(?:[0-9]+))))+)(?=$|[ ,;.)]))

and the two inputs are

a dispute under Section 3.1 (which shall be governed exclusively by Section 3.1) or as set forth in Section 11.3(b), the indemnification provisions of this Article XI and Article XII are the sole and exclusive remedies of the Parties pursuant to this Agreement or in connection with the transactions contemplated hereby. From and after the Closing, to the maximum extent permitted by Law, except with respect to claims based on intentional fraud, a dispute under Section 3.1 (which shall be governed exclusively by Section 3.1),

and

a dispute under Section 3.1 (which shall be governed exclusively by Section 3.1) or as set forth in Section 11.3(b), the indemnification provisions of this Article XI and Article XII are the sole and exclusive remedies of the Parties pursuant to this Agreement or in connection with the transactions contemplated hereby. From and after the Closing, to the maximum extent permitted by Law, except with respect to claims based on intentional fraud, a dispute under Section 3.1 (which shall be governed exclusively by Section 3.1),

also i am expecting

Section 3.1

Section 3.1

Section 11.3(b)

Article XI

Article XII

Section 3.1

Section 3.1

last one is not showing up in first input.

Answer 1

last one is not showing up in first input

One way of getting all characters is to consume up to an anchor character. For example take the text blah blah Section 3.1 (governed by Section 3.1) . We have three anchors, Section , ( and ) . Let us create a pattern based off of those literal anchors.

I now will comment my regex pattern which needs IgnorePattnerWhiteSpace to work properly in the regex parser btw.

(((Sub)?Section)|Article)\s+    # Anchor of Section or Article or Subsection
(?<Number>[^\s]+)               # Number involved
\s+
   \(                           # Anchor of '('
      (?<Conjuct>[^)]+)         # Consume til next anchor
   \)                           # ')' anchor.

By using the Not set [^ ] we can consume any funky characters which are not an ending ) anchor. Our match result looks like this

why are they returning different output.

You need to make the capturing less complex by maybe looking at the literal anchors as mentioned. Maybe even do a two pass regex by first creating general tokens of the text and upon the second regex pass, extract specific items from the tokens?

Also comment your pattern as I have done and work on individual pieces of it, once the individual items are working bring the whole pattern together.