Question

I have this regex that is supposed to match 1=’aa’ or 1=”aa” or 1=aa, and return number/value.

(?<number>\d)=\s*("|')?\s*(?<value>.*?)(?=("|')?\d=|$)

it works but it is returning the value incorrectly. Number comes back as 1 but value as aa” in the case of 1=”aa”

How can I get value = aa for the case of 1=”aa”.

Actual expression may contain 1=’aa’ 2=”bb” 3=cc etc.

Answer 1

The main problem here is that this:

(?=("|')?\d=|$)

means “followed by any of the following:

• " plus a digit plus =
• ' plus a digit plus =
• a digit plus =
• end-of-string

“. You’ll notice that it does not make any allowance for whitespace between the " or ' and the digit; so in the case of 1="aa" 2=..., the value simply is not allowed to be followed by " 2=.... Similarly, it does not make any allowance for " or ' plus end-of-string.

So the minimal fix is to allow — nay, require — some whitespace before the digit, and to move the ("|')? out of the lookahead assertion and into the main part of the regex:

("|')?(?=\s+\d=|$)

giving:

(?<number>\d)=\s*("|')?\s*(?<value>.*?)("|')?(?=\s+\d=|$)

While we’re at it, we might as well make some other tweaks to simplify the regex and reduce the number of cases where it can go wrong:

(?<number>\d)=\s*(["']?)(?<value>.*?)\1(?=\s+\d=|$)

(Further cleanup may be possible, but I don’t know enough about your data to recommend any more changes.)