decode this regex

Question

I've tried to understand the below but don't seem to get the last part of the regular expression which has {1,40}. Overall, I know the pattern tries to match the special characters and something else (the {1,40})

regexp_like(COLUMN,'^['||UNISTR('\0020')||'-'||UNISTR('\0060')||UNISTR('\007B')||UNISTR('\007D')||UNISTR('\007E')||UNISTR('\00C0')||'-'||UNISTR('\00DF')||']'||'{1,40}$')

It's a quantifier. The preceeding token (a character class here) will be repeated 1 to 40 times. — VLAZ, Jan 08 '20 at 20:27
'||' is putting me off which is right before {1,40}, and so I was not able to confirm my understanding. to me this seemed like an OR condition. So is this still a quantifier for the preceding character class? I am not trying to challenge, but I still have this doubt. — user3055262, Jan 08 '20 at 20:36
The `||` is the SQL string concatenation operator - it's not part of the regexp. — kfinity, Jan 08 '20 at 20:40
`||` is string concatenation in Oracle. `"[a-z]"||"{1,40}"` is equivalent to `"[a-b]" + "{1,40}"` in Java or `"[a-b]" . "{1,40}"` in PHP — VLAZ, Jan 08 '20 at 20:40
amazing. I fully understand it now. Thank you all for the quick resolution — user3055262, Jan 08 '20 at 20:58

GMB · Accepted Answer · 2020-01-08T21:08:21.060

regexp_like() checks that a string matches the regex provided as second argument.

Your regexp looks like ^[...]{1,40}$.

^ is the beginning of the string and $ is the end, so the entire string must match the regex.
[...] is a character class, that contains a bunch of characters code points. All characters of the string must belong to that list (any other character is forbiden). You would need to to check what they correspond to: unicode.org is your friend. For the first code points:

    \0020    space
    \0060    grave accent
    \007B    left curly bracket

finally, {1,40} is a quantifier: the length of the string must be at least one and at most 40.

decode this regex

1 Answers1