Simple Regex help for matching strings inside markup tags

[Greg Bernhardt]

Say I have a dozen of strings like this, with different content between the bold markup tags ("Re:" is also a constant). I need a regular expression that will match this. Basic and I think I am close, but just need a little help.



Here is what I have, but it's not matching

(?<=[b])Re:.*?(?=[/b]]) or this \[b]Re:(.+)[\/b]

 

[WWGD]

Say I have a dozen of strings like this, with different content between the bold markup tags ("Re:" is also a constant). I need a regular expression that will match this. Basic and I think I am close, but just need a little help.



Here is what I have, but it's not matching

(?<=[b])Re:.*?(?=[/b]]) or this \[b]Re:(.+)[\/b]

Is this Python?

 

[Greg Bernhardt]

Just vanilla regular expression to use in SQL eventually

 

[Ibix]

Don't you just need to escape the square brackets? Otherwise the regex interprets it as a character range (i.e., [abcd] matches a, b, c or d, so a b in square brackets matches b). So \[b\]Re:.*\[\/b\] should work (although that would be case sensitive).

Edit: And whether you need to escape the slash in the close bold tag may depend on your language.

 

[PeterDonis]

Try this:

\[b\]Re:(.+)\[/b\]

This assumes that what's after the "Re:" can be anything; if you want to restrict it to alphanumerics and punctuation you could do something like:

\[b\]Re:(([A-Za-z \!\?\.\,\;]+))\[/b\]

 

[WWGD]

So PF comment from the admin side is SQL-queryable?

 

[PeterDonis]

whether you need to escape the slash in the close bold tag may depend on your language.

Python's regexp parser (which is what I used for my testing) doesn't seem to require it, but yes, others might.

 

[Greg Bernhardt]

Try this:

\[b\]Re:(.+)\[/b\]

This assumes that what's after the "Re:" can be anything; if you want to restrict it to alphanumerics and punctuation you could do something like:

\[b\]Re:(([A-Za-z \!\?\.\,\;]+))\[/b\]

This works when testing in https://regexr.com/ but when I do a SELECT query using it, it's not at all working and returns results that don't match at all. hmmm

So PF comment from the admin side is SQL-queryable?

Of course, it's all stored in a DB

 

[PeterDonis]

when I do a SELECT query using it, it's not at all working and returns results that don't match at all

What database engine? And what exactly is the SELECT query statement?

 

[Greg Bernhardt]

What database engine? And what exactly is the SELECT query statement?

MariaDB which is a drop in replacement for mySQL.

SELECT * FROM `xf_post` WHERE `message` REGEXP '\[b\]Re:(.+)\[\/b\]'

This actually returns only 6 results when I know there are 127k when doing a LIKE query for "[B]Re:"

One of the results matches this

[b]Please check out the COBE and WMAP results[/b]

 

[WWGD]

This works when testing in https://regexr.com/ but when I do a SELECT query using it, it's not at all working and returns results that don't match at all. hmmm
Of course, it's all stored in a DB

In case you're interested, Sql Server dev allows for Python and has its own ML Server.

 

[PeterDonis]

MariaDB

From the MariaDB docs online, it looks like you actually need two backslashes to properly escape characters:

https://mariadb.com/kb/en/library/regular-expressions-overview/#escaping

 

[Greg Bernhardt]

From the MariaDB docs online, it looks like you actually need two backslashes to properly escape characters:

https://mariadb.com/kb/en/library/regular-expressions-overview/#escaping

I think that is bingo! I changed to \\[b\]Re:(.+)\\[\\/b\] and think it looks like a good result!

 

[Ibix]

Python's regexp parser (which is what I used for my testing) doesn't seem to require it, but yes, others might.

I tested by piping a string through sed, which does require it because it uses / to delimit the regex. Some other engines follow suit - apparently not this one, though.

 

[PeterDonis]

I think that is bingo!