Regex Isn't Hard (2023)

In a previous job I've done some stupid tricks with regexes. Inside a MongoDB database I had documents with a version field in string form ("x.y.z") and I needed to exclude documents with a schema too old to process in my queries.

One can construct a regex that matches a number between x and y by enumerating all the digit patterns that fit the criteria. For example, the following pattern matches a number between 1 and 255: ^([1-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-5])$

This can be extended to match a version less than or equal to x.z.y by enumerating all the patterns across the different fields. The following pattern matches any version less than or equal to 2.14.0: ^([0-1]\.\d+\.\d+)|(2\.[0-9]\.\d+|(2\.1[0-3]\.\d+))$

Basically, I wrote a Java method that would generate a regex with all the patterns to match a version greater than or equal to a lower bound, which was then fed to MongoDB queries to exclude documents too old to process based on the version field. It was a stupid solution to a dumb problem, but it worked flawlessly.