Make sure the presence of a particular character in a string - regex

Question

I'm using the following regex to parse my application log file to search for particular string
\[\s*\b(?:[0-9A-Za-z][0-9A-Za-z-_.#]{0,200})(?:\.(?:[0-9*A-Za-z][0-9A-Za-z-_.#]{0,200}))*(\.?|\b)\s*]
This works fine but now we need to make sure that the string "must" contain "-" character to match. I'm confused to add this condition to the original regx.
Any pointers would be helpful.
Thanks and Regards,
Santhosh

Answer 1

The regex matches a string inside square brackets, [ and ], and may only consist of non-[ and non-] symbols.
You can easily add a positive lookahead restriction after the opening [ like check if the next characters other than ] and [ are followed with -:
\[ # opening [
(?=[^\]\[]*-) # There must be a hyphen in [...]
\s* # 0+ whitespaces
\b(?:[0-9A-Za-z][0-9A-Za-z-_.#]{0,200}) # Part 1 (with obligatory subpattern)
(?:\. # Part 2, optional
(?:[0-9*A-Za-z][0-9A-Za-z-_.#]{0,200})
)*
(\.?|\b) # optional . or word boundary
\s* # 0+ whitespaces
] # closing ]
See the regex demo
And a one-liner:
\[(?=[^\]\[]*-)\s*\b(?:[0-9A-Za-z][0-9A-Za-z-_.#]{0,200})(?:\.(?:[0-9*A-Za-z][0-9A-Za-z-_.#]{0,200}))*(\.?|\b)\s*]
Tip: use the verbose /x modifier to split the pattern into separate multiline blocks for analysis, it will help you in the future when you need to modify the pattern again.
If you need to match only if - or # is present inside [...], modify the lookahead as (?=[^\]\[]*[-#]). For a more general case, use (?=[^\]\[]*(?:one|another|must-be-present)) alternatives inside an additional group inside the lookahead.

Answer 2

Updated answer - Assertion
In this case, the better way to do it is to use an assertion consisting of checking only the position's expected to match the character in question.
I know it's simple, but using the outter pseudo-anchor text \[ ... \] as
a delimiter that cannot exist in the body is a rarity.
You should always try to avoid doing it like this.
Things change, your input could change.
The rule to follow in validation of known characters that are Mid-String is to use only them
when using an assertion validator.
This avoids the necessity of relying on what is not there at the moment ie, not a ],
but should rely on what is there.
Again, this pertains to mid-string matching.
BOL/EOL is a different thing entirely ^$, and is a more permanent construct
with which to leverage.
It's always better to code smarter.
\[\s*\b(?=[0-9A-Za-z][0-9A-Za-z_.#]{0,199}-|[0-9A-Za-z][0-9A-Za-z_.#]{0,200}(?:\.[0-9*A-Za-z][0-9A-Za-z_.#]{0,200})*\.[0-9*A-Za-z][0-9A-Za-z_.#]{0,199}-)(?:[0-9A-Za-z][0-9A-Za-z_.#-]{0,200})(?:\.(?:[0-9*A-Za-z][0-9A-Za-z_.#-]{0,200}))*(\.?|\b)\s*\]
Using Conditionals
If your engine supports conditionals, the easy way is to not rely on a fluke
of pseudo anchor text, ie. [..].
\[\s*\b[0-9A-Za-z](?:[0-9A-Za-z_.#]|(-)){0,200}(?:\.(?:[0-9*A-Za-z](?:[0-9A-Za-z_.#]|(-)){0,200}))*(\.?|\b)\s*\](?(1)|(?(2)|(?!)))
Expanded
\[ \s* \b
[0-9A-Za-z]
(?:
[0-9A-Za-z_.#]
| ( - ) # (1)
){0,200}
(?:
\.
(?:
[0-9*A-Za-z]
(?:
[0-9A-Za-z_.#]
| ( - ) # (2)
){0,200}
)
)*
( \.? | \b ) # (3)
\s* \]
(?(1) # Fail if no dash found
| (?(2)
| (?!)
)
)

Answer 3

This conditional would work if just want to make sure that - occurs within your string before running your block of code.
if (myString.indexOf('-') >= 0) {
//your code
}

Answer 4

If you have to have a single hyphen, you'll have to either repeat most of the pattern, or check for it in a second phase:
if re.match(pattern, line):
if not '-' in line:
raise MissingDash('No dash in line: {}'.format(line))
I'd suggest adding the second check, since adding the requirement to the regex would make it even more horrible to read.