I currently have the following regular expression:
/(_|[a-z]|[A-Z])(_|[a-z]|[A-Z]|[0-9])*/
I would like the expression not to match with "PI", however I failed to do so. To clarify, I would like the following to be valid: _PI, abcPI, PIpipipi I just dont want to accept PI when its on its own.
Before jumping at the solution, please have a look at your regex: the character classes for single ranges inside alternation groups is an inefficient way of writing regex patterns. You may simply merge these ([A-Z]|[0-9]|_)+ into [A-Z0-9_]+.
The solution may be a word boundary with a negative lookahead after it:
r"\b(?!PI\b)[_a-zA-Z][_a-zA-Z0-9]*"
See the regex demo. You may replace [a-zA-Z0-9_] with \w:
re.compile(r"\b(?!PI\b)[_a-zA-Z]\w*") # In Python 2.x, re.UNICODE is not enabled by default
re.compile(r"\b(?!PI\b)[_a-zA-Z]\w*", re.A) # In Python 3.x, make \w match ASCII only
Details
\b - word boundary(?!PI\b) - immediately to the right, there can't be PI as a whole word[_a-zA-Z] - an ASCII letter or _[_a-zA-Z0-9]* - 0 or more underscores, ASCII letters or digits.Submitting another answer:
^(((?!PI).)*)$|^.*(PI).+$|^.+(PI).*$
I broke it down into 3 cases using OR |:
1) Match a string that doesn't contain PI at all.
^(((?!PI).)*)$
2) Match a string that has PI in it but has at least one character behind it, and optionally any characters ahead of it.
^.*(PI).+$
3) Match a string that has PI in it but has at least one character ahead of it, and optionally any characters behind it.
^.+(PI).*$
Here it is with test cases: https://regex101.com/r/7rzqpe/3
Please comment if you find a missing edge case.
Not very nice, but I'll add it anyway for variety:
/^([A-OQ-Za-z_][A-Za-z0-9_]*|P([A-HJ-Za-z0-9_][A-Za-z0-9_]*)?)$/
PIthat you want to exclude?r"\b(?!PI\b)[_a-zA-Z][_a-zA-Z0-9]*"? See demo.