[hibernate-dev] SQM-11 : Consider a new matches operator - thoughts?

[Steve Ebersole]

My phone is being crazy with sending replies :(

So if we don't want to use ~ here, I am fine with that.  However I do not
want this used for full text searching.  Your argument is that its "not a
natural fit" for use in this nullness checking, but that same argument fits
full text searching.

On Sat, Aug 29, 2015 at 6:17 PM Steve Ebersole <steve at hibernate.org> wrote:

> So if we don't want to use ~ here, I am fine with that.  However I do not
> want this used for full text searching.
>
> On Sat, Aug 29, 2015, 6:07 PM Sanne Grinovero <sanne at hibernate.org> wrote:
>
>> I think we all agree that we want at least the high level intent of an
>> HQL query to be understandable without having to keep a reference
>> sheet of its special symbols at hand.
>>
>> I would interpret that "~" symbol differently, so I don't think it's
>> suited to imply some special meaning related to "null" and while I
>> normally understand we argue about "naming things" as we all have
>> different language backgrounds, this one seems universally established
>> in the scientific world, so I'm afraid that it would look a really odd
>> choice to most users.
>>
>> Gunnar's proposal looks better, as that's commonly used in similar
>> scenarios: beyond Groovy, it's commonly used for ternary operator
>> which is essentially a short hand for a conditional:
>>   - https://en.wikipedia.org/wiki/%3F:
>>
>> On full-text query, looks like a clarification is in order on why I
>> was thinking about that symbol specifically:
>> A full-text query goes far beyond "containing" some keywords; for
>> example it can provide matches on synonyms, disregard typos, expand
>> acronyms, or even give different semantic value for words depending on
>> their position and role in a sentence. The match is really about
>> "search stuff which is semantically similar to this".
>> I realize that most SQL technologies provide only a limited subset of
>> such options, and that's exactly why we provide alternatives like
>> Hibernate Search which do much more and better ;-)
>> i.e. with Search such an operator would mean "equals or almost
>> equals", for some definition of "almost equals" which we allow the
>> user to specify precisely in the mapping and configuration.
>>
>> Thanks,
>> Sanne
>>
>> On 29 August 2015 at 18:21, Steve Ebersole <steve at hibernate.org> wrote:
>> > Sorry, but I agree with Brett.  CONTAINS is a much more natural way to
>> > perform a full-text search, especially for those coming from SQL.
>> >
>> > On Sat, Aug 29, 2015 at 1:47 AM Gunnar Morling <gunnar at hibernate.org>
>> wrote:
>> >
>> >> I like the idea of a new operator, but I side with Sanne that "~"
>> >> would be useful for similarity/full-text searches.
>> >>
>> >> What about "?=", somewhat inspired by Groovy's Elvis operator? It
>> >> seems nicer to me for expressing null-awareness. The pattern could
>> >> even be generalized by using the operator prefix for making some more
>> >> comparison operators null-aware: "?<", "?>" etc.
>> >>
>> >> --Gunnar
>> >>
>> >>
>> >>
>> >> 2015-08-28 16:36 GMT+02:00 Brett Meyer <brett at hibernate.org>:
>> >> > +1 to ~.  Sanne has a good point, but I think I'd rather see a
>> function
>> >> > name there (CONTAINS, etc.).
>> >> >
>> >> > On 08/28/2015 10:09 AM, Sanne Grinovero wrote:
>> >> >> On 28 August 2015 at 15:02, Steve Ebersole <steve at hibernate.org>
>> wrote:
>> >> >>> What do y'all think of using a symbol like ~ for this?  The idea
>> would
>> >> be
>> >> >>> similar to the "wavy equals" from logic used to denote
>> "approximately
>> >> >>> equals".
>> >> >> I was hoping that one day we would be able to use the '~' symbol for
>> >> >> full-text queries, i.e. a fuzzy match for text fields to extend HQL.
>> >> >> But we have no concrete plans about that, and we currently don't do
>> a
>> >> >> great job in allowing people to combine full-text restrictions with
>> >> >> relational restrictions, so that might be an unrealistic dream.
>> >> >>
>> >> >>
>> >> >>> On Thu, Aug 27, 2015 at 7:05 AM andrea boriero <
>> andrea at hibernate.org>
>> >> wrote:
>> >> >>>
>> >> >>>> i like the idea of "matches" operator for dealing with "is null".
>> >> >>>> +1
>> >> >>>>
>> >> >>>> On 26 August 2015 at 19:32, Steve Ebersole <steve at hibernate.org>
>> >> wrote:
>> >> >>>>
>> >> >>>>> https://hibernate.atlassian.net/browse/SQM-11
>> >> >>>>> _______________________________________________
>> >> >>>>> hibernate-dev mailing list
>> >> >>>>> hibernate-dev at lists.jboss.org
>> >> >>>>> https://lists.jboss.org/mailman/listinfo/hibernate-dev
>> >> >>>>>
>> >> >>>>
>> >> >>> _______________________________________________
>> >> >>> hibernate-dev mailing list
>> >> >>> hibernate-dev at lists.jboss.org
>> >> >>> https://lists.jboss.org/mailman/listinfo/hibernate-dev
>> >> >> _______________________________________________
>> >> >> hibernate-dev mailing list
>> >> >> hibernate-dev at lists.jboss.org
>> >> >> https://lists.jboss.org/mailman/listinfo/hibernate-dev
>> >> >
>> >> >
>> >> > _______________________________________________
>> >> > hibernate-dev mailing list
>> >> > hibernate-dev at lists.jboss.org
>> >> > https://lists.jboss.org/mailman/listinfo/hibernate-dev
>> >> _______________________________________________
>> >> hibernate-dev mailing list
>> >> hibernate-dev at lists.jboss.org
>> >> https://lists.jboss.org/mailman/listinfo/hibernate-dev
>> >>
>> > _______________________________________________
>> > hibernate-dev mailing list
>> > hibernate-dev at lists.jboss.org
>> > https://lists.jboss.org/mailman/listinfo/hibernate-dev
>> _______________________________________________
>> hibernate-dev mailing list
>> hibernate-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/hibernate-dev
>>
>