Question

I am using a regular expression to match where conditions in a SQL query.

I want WHERE <ANY CONDITION>, but with the exception of WHERE ROWNUM <WHATEVER>.
So I do not want ROWNUM to appear after the WHERE keyword.

I did use Lookaheads to achieve that. My regex is WHERE (.*(?! ROWNUM )+). The problem is, it still matches WHERE ROWNUM < 1000. If I delete the space before ROWNUM in the regex, then any column with a name ending with ROWNUM won’t match. If I delete the space after WHERE then it would match even if there is no space after the WHERE keyword. However, if there are two spaces or any other character between ROWNUM and the WHERE keyword (might be a condition), then it is ok. So if ROWNUM is first in the condition my regex does not work.

How can I fix this ?

Answer 1

It sounds like you want

WHERE(?!.*\bROWNUM\b).*

which will match WHERE .*, unless the .* contains a ROWNUM that is surrounded by word boundaries. (\b, “word boundary”, is a zero-width assertion denoting a position that is preceded by a letter or digit or underscore or followed by a letter or digit or underscore, but not both.)