On 11-2-2019 23:51, [email protected] [firebird-support] 
wrote:
> I believe this has been done with such database engines as SQL Server 
> and PostgreSQL.
> 
> However, the Firebird Database Engine has a stricter standard as to what 
> is implemented in the SQL language than other such engines.  
> Nonetheless, it appears that there is some similar capabilities as the 
> following link describes... 
> https://stackoverflow.com/questions/21432507/repetitions-with-regexes-in-firebird

That doesn't have to do with a 'stricter standard', it has to do with 
having a smaller team and having to choose what you do and don't. And in 
that case, following the ISO-9075 as a guideline instead of inventing 
your own, is a better solution.

But, that said, the SQL standard actually describes more extensive 
regular expression features than Firebird currently has (eg LIKE_REGEX, 
OCCURRENCES_REGEX, POSITION_REGEX, SUBSTRING_REGEX, TRANSLATE_REGEX), 
which btw don't use the atrocity that is the SQL regex used by SIMILAR 
TO and SUBSTRING(.. SIMILAR TO ...), but the XSLT variant which is 
similar to Java's and Perl. This also provide support for groups, etc.

> It may be more convenient to pull out data via a Regex expression in SQL 
> code but it is rather redundant since all of the major development 
> languages already support extensive Regex pattern matching capabilities...

It can have its uses where retrieving data out to a programming language 
and then back again to the database is inefficient. But I agree, in 
general, a need for this would be symptomatic of having individual data 
items encoded in a single string.

Mark
-- 
Mark Rotteveel

Reply via email to