I'm trying to use a regex to detect URLs in all the rows of my table, here's the regex
\b(([\w-]+:\/\/?|www[.])[^\s()<>]+(?:\([\w\d]+\)|([^[:punct:]\s]|\/)))
However, I invariably get the "repetition-operator operand invalid" error, which, after hours of search on the internet, still remains obscure.
Where have I gone wrong? What can I do to fix this? And alternaltively, is there a better way to detect URLs in messages in SQL other than a Regex?
Thank you.
You cannot use ? quantifier in MySQL regex as the syntax is POSIX-based. Still, you can use * to match 0 or more characters. Also, \b in MySQL regex should be replaced with [[:<:]] (since this matches at the beginning of a word).
Thus, I suggest using
[[:<:]](([a-zA-Z0-9-]+:\/\/*|www[.])[^ ()<>]+(\([a-zA-Z0-9_]+\)|([^ [:punct:]]|\/)))
I am expanding \w to [a-zA-Z0-9_] as it is exactly what \w is. Instead of \s, I am using a literal space. Instead of \d, I am using [0-9]. This is done for readability and better compatibility. If \w, \d and \s work for you, you can use them, but I do not see them among the supported entities in POSIX specs.
Also, instead of literal space, you could use [:space:], it matches space, tab, newline, and carriage return. Instead of [a-zA-Z] you can use [:alpha:], and instead of [0-9], you can use [:digit:]. Please also check this:
[[:<:]](([[:alpha:][:digit:]-]+:\/\/*|www[.])[^[:space:]()<>]+(\([[:alpha:][:digit:]_]+\)|([^[:space:][:punct:]]|\/)))
Related
I'm using regex word boundary \b, and I'm trying to match a word in the following sentence but the result is not what I need. Connector Punctuations (such as underscore) are not being considered as a word boundary
Sentence: ab﹎cd_de_gf|ij|kl|mn|op_
Regexp: \\bkl\\b
However, de is not getting matched.
I tried updating the regexp to use unicode connector punctuation (it's product requirement as we support CJK languages as well) but that isn't working.
Regexp: (?<=\\b|[\p{Pc}])de(?=\\b|[\p{Pc}])
What am i missing here?
Note: (?<=\\b|_)de(?=\\b|_) seems to work for underscores but i need the regex to work for all the connector punctuations.
Thanks in advance !!
Based on the use case you have described you can simplify your regex to:
(?<![[:alnum:]])de(?![[:alnum:]])
instead of trying to match word boundaries, unicode punctuation characters etc.
This will match de if it not followed or preceded by any alpha-numeric character.
To match any connector punctuation characters you need \p{Pc}:
(?<=\\b|\\p{Pc})de(?=\\b|\\p{Pc})
NOTE: \p{Pc} can also be written as [_\u203F\u2040\u2054\uFE33\uFE34\uFE4D-\uFE4F\uFF3F] that matches all these 10 chars.
While there's a told of "old" examples on the internet using the now unsupported '[[:<:]]word[[:>:]]' technique, I'm trying to find out how, in MySQL 8.0.30, to do exact word matching from our table with words that have special characters in them.
For example, we have a paragraph of text like:
"Senior software engineer and C++ developer with Unit Test and JavaScript experience. I also have .NET experience!"
We have a table of keywords to match against this and have been using the basic system of:
SELECT
sk.ID
FROM
sit_keyword sk
WHERE
var_text REGEXP CONCAT('\\b',sk.keyword,'\\b')
It works fine 90% of the time, but it completely fails on:
C#, C++, .NET, A+ or "A +" etc. So it's failing to match keywords with special characters in them.
I can't seem to find any recent documentation on how to address this since, as mentioned, nearly all of the examples I can find use the old unsupported techniques. Note I need to match these words (with special characters) anywhere in the source text, so it can be the first or last word, or somewhere in the middle.
Any advice on the best way to do this using REGEXP would be appreciated.
You need to escape special chars in the search phrase and use the construct that I call "adaptive dynamic word boundaries" instead of word boundaries:
var_text REGEXP CONCAT('(?!\\B\\w)',REGEXP_REPLACE(sk.keyword, '([-.^$*+?()\\[\\]{}\\\\|])', '\\$1'),'(?<!\\w\\B)')
The REGEXP_REPLACE(sk.keyword, '([-.^$*+?()\\[\\]{}\\\\|])', '\\$1') matches . ^ $ * + - ? ( ) [ ] { } \ | chars (adds a \ before them) and (?!\\B\\w) / (?<!\\w\\B) require word boundaries only when the search phrase start/ends with a word char.
More details on adaptive dynamic word boundaries and demo in my YT video.
Regular expressions treat several characters as metacharacters. These are documented in the manual on regular expression syntax: https://dev.mysql.com/doc/refman/8.0/en/regexp.html#regexp-syntax
If you need a metacharacter to be treated as the literal character, you need to escape it with a backslash.
This gets very complex. If you just want to search for substrings, perhaps you should just use LOCATE():
WHERE LOCATE(sk.keyword, var_text) > 0
This avoids all the trickery with metacharacters. It treats the string of sk.keyword as containing only literal characters.
python pattern => ^(?=.\bABDUL\b)(?=.\bHAI\b.)(?=.\bMANSOOR\b).*$
need equalent mysql pattern
can you please help me out ?
The regex in question is a quite strange way how to match simple words. It is not clear what is the expected input. Maybe, the input justifies this approach.
^(?=.\bABDUL\b)(?=.\bHAI\b.)(?=.\bMANSOOR\b).*$
Which means: At the beginning there must be any character which is not a part of a word, then ABDUL, a non word character, HAI, a non word character, MANSOOR, a non word character or the end of the string.
^[^[:alnum:]]ABDUL[^[:alnum:]]HAI[^[:alnum:]]MANSOOR([^[:alnum:]]?.*)?$
Which is: At the beginning, not a number or alphabet character (alphanumerical), ABDUL, one non-alphanumerical, HAI, one non-alphanumerical, MANSOOR one non-alphanumerical or the end of the string.
I did not test it and did not intended to make it 100% the same as the first one, but it should be close enough.
For anyone who would like to copy it to their code:
Matching the first character is not very common and can be a bug in the original regexp.
(?=...) is an "lookahead assertion" which does not consume any characters, the POSIX version does not have it, but for a simple string searching it may not be important.
Both versions should match strings like !ABDUL$HAI)MANSOOR - make sure that this is what you want.
For someone who would like to understand the regular expressions I used
https://dev.mysql.com/doc/refman/8.0/en/regexp.html for mysql (POSIX syntax) and https://docs.python.org/3/library/re.html for python (PCRE = Perl compatible syntax)
Let's say I'm looking for a specific number in string with newlines:
1\n2\n4\n5\n7\n8\n9\n12\n13
Lookahead and lookbehind works perfectly with something like:
(?<![0-9])12(?![0-9])
See Regex 101 demo.
However, I need a workaround for the MySQL REGEXP, as it does not support patterns with lookarounds.
If (?<![0-9])12(?![0-9]) works for you in an online regex engine, the following regex must work for you in REGEXP pattern that is POSIX based:
(^|[^0-9])12($|[^0-9])
Or, you may use word boundaries:
[[:<:]]12[[:>:]]
See the regex demo.
Details:
(^|[^0-9]) - start of string or a non-digit char
[[:<:]] - leading word boundary
12 - a specific numeric value, 12 here
[[:>:]] - trailing word boundary
($|[^0-9]) - end of string or a non-digit.
For MySQL 8.0,
"\\b12\\b"
or
"\\s12\\s"
The former checks for "word boundaries" and might be better. The latter checks for whitespace, such as \n.
The doubled backslashes may be overkill, depending on the client.
I'm trying to recognize quoting (citing) somebody's else sentence in a markdown text, which I have in my local copy of MySQL GHTorrent dataset. So I wrote this query:
select * from github_discussions where body rlike '(.)*(\s){1,}(>)(\s){1,}(.)+';
it matches some unwanted data, which according to https://regex101.com/, it should not with this particular regular expression.
Test string:
`Params` is plural -> contain<s>s</s>
Matched on MySQL database, not matched at regex101 dot com.
Obvious example of quoting, but not matched at db:
Yes, I believe so.\r\n\r\n\r\n\r\nK\r\n\r\n> On 19-Jul-2014, at 17:33, Stefan Karpinski <notifications#github.com> wrote:\r\n> \r\n> This is the standard 3-clause BSD license, right?\r\n> \r\n> —\r\n> Reply to this email directly or view it on GitHub.
Moreover, MySQL workbench didn't show those return carriage and new line symbols unless copy-pasted here.
Can I normalize (remove \r and \n) with some update query ?
Is MySQL regex implementation different from POSIX standard regex ?
Do you have by any chances maximally clean solution for recognizing quoting in a markdown text ?
Thanks!
You've got an awful lot of parens in there. Try this as functionally what you have above:
select * from github_discussions where body rlike '.*[:blank:]+>[:blank:]+.+'
However, I'm not sure that's really what you want. This would happily match this line:
this is before > and after
which by my understanding is not a quoted string in markdown. Instead I would anchor it at the beginning like this:
select * from github_discussions where body rlike '^[:blank:]*>[:blank:]+'
That will match a greater-than sign at the beginning of the line, optionally preceded by whitespace. Is that what you are looking for?
I'm not sure if your data has newlines embedded. If so, you may need to look into ways of having your regex identify newlines using the ^ anchoring symbol. As is the well accepted conclusion in regex literature, that is left as an exercise for the student. :-)