Luigi,

It's been almost two months since we asked about this.  We'll turn off the
strictSQL option if we have to, but that doesn't seem like a good long-term
solution.  We really would like a way to do this.

We're having major memory issues with our current version of OrientDB, and
would like to upgrade, but this is really standing in the way.

Please let us know what we can do.  We've been patient, but this seems
extraordinary.

Steve

On Sun, Apr 8, 2018 at 11:38 PM, Steven Tomer <steven.to...@gmail.com>
wrote:

> Luigi,
>
> When is the release of 3.0.0 planned?
>
> We're still waiting to hear back from you on how to do this.
>
> Thanks,
>
> Steve
>
> On Mon, Mar 26, 2018, 8:45 AM Luigi Dell'Aquila <
> luigi.dellaqu...@gmail.com> wrote:
>
>> Hi Steve,
>>
>> Sorry, we are pretty busy with the release of v 3.0.0, I'll try to follow
>> up as soon as I can.
>>
>> The traverse() executor is still there, but only in the old SQL executor.
>> The new parser (strict SQL) was implemented for the new executor, so it
>> doesn't make much sense to have it in the parser without implementing it in
>> the new executor.
>> I have to say that in general we don't like the traverse() operator
>> syntax much, and now we have many alternatives like TRAVERSE and MATCH
>> queries that are supposed to do the same thing, so adding complexity
>> re-introducing a legacy behavior does not sound a very good choice.
>> To be honest, since we switched on the strictSQL option by default (more
>> than two years ago), it's maybe the second time we have an explicit request
>> for the traverse() operator, and in both cases it was for a migration from
>> previous versions, so I'm pretty confident that we can cover your use case
>> what we have, with better performance and a more elegant way.
>>
>> I'll keep you posted
>>
>> Thanks
>>
>> Luigi
>>
>>
>> 2018-03-26 15:52 GMT+02:00 Steven Tomer <steven.to...@gmail.com>:
>>
>>> Luigi,
>>>
>>> I haven't heard back on this.
>>>
>>> Since the backend to the traverse() is still there, we're wondering how
>>> much trouble it would be to add it into the new parser.
>>>
>>> It definitely has some powerful and unique use cases.  If these aren't
>>> covered by the new match projection, I would suggest that it should live on
>>> to cover them.
>>>
>>> Thanks,
>>>
>>> Steve
>>>
>>> On Tue, Mar 20, 2018, 9:37 AM Steven Tomer <steven.to...@gmail.com>
>>> wrote:
>>>
>>>> Thanks so much!
>>>>
>>>> Steve
>>>>
>>>> On Tue, Mar 20, 2018 at 9:34 AM, Luigi Dell'Aquila <
>>>> luigi.dellaqu...@gmail.com> wrote:
>>>>
>>>>> Hi Steve,
>>>>>
>>>>> Thank you very much, I'll check it asap.
>>>>> There is a way to re-enable the old traverse, you have to disable
>>>>> strict SQL parsing:
>>>>>
>>>>> ALTER DATABASE CUSTOM strictSQL=false
>>>>>
>>>>> Anyway, I wouldn't recommend it in the long run, as you will lose a
>>>>> lot of stability in SQL parsing
>>>>>
>>>>> I'll give you a feedback asap about the query
>>>>>
>>>>> Thanks
>>>>>
>>>>> Luigi
>>>>>
>>>>> 2018-03-20 16:30 GMT+01:00 Steven Tomer <steven.to...@gmail.com>:
>>>>>
>>>>>> Luigi,
>>>>>>
>>>>>> I've zipped up the dataset I've been using and am attaching it here.
>>>>>>
>>>>>> We really need a way to do queries like this.  Is there any chance we
>>>>>> can re-enable the old traverse() operator?
>>>>>>
>>>>>> Thanks,
>>>>>>
>>>>>> Steve
>>>>>>
>>>>>> On Tue, Mar 13, 2018 at 9:08 AM, Luigi Dell'Aquila <
>>>>>> luigi.dellaqu...@gmail.com> wrote:
>>>>>>
>>>>>>> Hi Steve,
>>>>>>>
>>>>>>> ah, got it, the old traverse() stopped when it found duplicates,
>>>>>>> while the MATCH doesn't.
>>>>>>>
>>>>>>> Do you have a dataset I can check, so that I can try it and give you
>>>>>>> better help?
>>>>>>>
>>>>>>> Thanks
>>>>>>>
>>>>>>> Luigi
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> 2018-03-13 16:03 GMT+01:00 Steven Tomer <steven.to...@gmail.com>:
>>>>>>>
>>>>>>>> Luigi,
>>>>>>>>
>>>>>>>> That query never returns (looks like it goes into an infinite loop).
>>>>>>>>
>>>>>>>> I had to kill the server process to make it stop.
>>>>>>>>
>>>>>>>> Any other ideas?
>>>>>>>>
>>>>>>>> Steve
>>>>>>>>
>>>>>>>> On Tue, Mar 13, 2018 at 8:16 AM, Luigi Dell'Aquila <
>>>>>>>> luigi.dellaqu...@gmail.com> wrote:
>>>>>>>>
>>>>>>>>> Hi Steve,
>>>>>>>>>
>>>>>>>>> Try this
>>>>>>>>>
>>>>>>>>> SELECT expand(person) FROM (
>>>>>>>>>   MATCH {class: Person, as: person, where: (givenname =
>>>>>>>>> 'William')}.in(){while: ($depth = 0 OR givenname != 'George'), where:
>>>>>>>>> (givenname = 'George')} RETURN person
>>>>>>>>> )
>>>>>>>>>
>>>>>>>>> The $depth = 0 guarantees that you don't stop traversing
>>>>>>>>> immediately; the givenname != 'George' just stops traversing as
>>>>>>>>> soon as you find a "George", I guess it is what you want (you don't 
>>>>>>>>> need
>>>>>>>>> duplicates)
>>>>>>>>>
>>>>>>>>> Thanks
>>>>>>>>>
>>>>>>>>> Luigi
>>>>>>>>>
>>>>>>>>> 2018-03-13 14:52 GMT+01:00 Steven Tomer <steven.to...@gmail.com>:
>>>>>>>>>
>>>>>>>>>> Hello Luigi,
>>>>>>>>>>
>>>>>>>>>> That's sad news indeed.  It will require a major task to rip out
>>>>>>>>>> all of the places we're using traverse().
>>>>>>>>>>
>>>>>>>>>> I've been playing around with the MATCH operator, and I can't get
>>>>>>>>>> it to model the former query:
>>>>>>>>>>
>>>>>>>>>> SELECT FROM Person WHERE givenname = 'William' and in
>>>>>>>>>> traverse(0,-1,'in') ( givenname = 'George' )
>>>>>>>>>>
>>>>>>>>>> I've tried:
>>>>>>>>>>
>>>>>>>>>> MATCH {class: Person, as: person, where: (givenname =
>>>>>>>>>> 'William')}.in(){while: ($depth > 0), where: (givenname = 'George')} 
>>>>>>>>>> RETURN
>>>>>>>>>> person
>>>>>>>>>> MATCH {class: Person, as: person, where: (givenname =
>>>>>>>>>> 'William')}.in('in'){while: ($depth > 0), where: (givenname = 
>>>>>>>>>> 'George')}
>>>>>>>>>> RETURN person
>>>>>>>>>> MATCH {class: Person, as: person, where: (givenname =
>>>>>>>>>> 'William')}.in('MyEdge'){while: ($depth > 0), where: (givenname
>>>>>>>>>> = 'George')} RETURN person
>>>>>>>>>>
>>>>>>>>>> None of your examples have an intermediary 'edge' type between
>>>>>>>>>> them.  We have Person -> MyEdge -> Person -> MyEdge -> Person.  The 
>>>>>>>>>> traverse
>>>>>>>>>> operator lets me hop over the edges.  How do you do it with MATCH?
>>>>>>>>>>
>>>>>>>>>> Steve
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> On Tue, Mar 13, 2018 at 2:07 AM, Luigi Dell'Aquila <
>>>>>>>>>> luigi.dellaqu...@gmail.com> wrote:
>>>>>>>>>>
>>>>>>>>>>> Hi Steven,
>>>>>>>>>>>
>>>>>>>>>>> The traverse() operator was deprecated long time ago and is not
>>>>>>>>>>> fully supported anymore.
>>>>>>>>>>> I strongly suggest you to use TRAVERSE queries (
>>>>>>>>>>> https://orientdb.com/docs/2.2.x/SQL-Traverse.html) or even
>>>>>>>>>>> better MATCH queries (https://orientdb.com/docs/2.
>>>>>>>>>>> 2.x/SQL-Match.html)
>>>>>>>>>>>
>>>>>>>>>>> Another thing: if you are migrating now, you could probably
>>>>>>>>>>> consider v 3.0 (see https://orientdb.com/orientdb-labs/), the
>>>>>>>>>>> final GA will be released this week
>>>>>>>>>>>
>>>>>>>>>>> Thanks
>>>>>>>>>>>
>>>>>>>>>>> Luigi
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> 2018-03-12 23:00 GMT+01:00 StevenTomer <steven.to...@gmail.com>:
>>>>>>>>>>>
>>>>>>>>>>>> Hello,
>>>>>>>>>>>>
>>>>>>>>>>>> I'm using OrientDB 2.2.32.  We're updating our software from
>>>>>>>>>>>> 2.0.8.
>>>>>>>>>>>>
>>>>>>>>>>>> The traverse() operator in selects does not appear to be
>>>>>>>>>>>> working.
>>>>>>>>>>>>
>>>>>>>>>>>> When I run the attached program, I get a com.orientechnologies.
>>>>>>>>>>>> orient.core.sql.OCommandSQLParsingException.
>>>>>>>>>>>>
>>>>>>>>>>>> Exception in thread "main" com.orientechnologies.orient.
>>>>>>>>>>>> core.sql.OCommandSQLParsingException: Error parsing query:
>>>>>>>>>>>> SELECT FROM Person WHERE givenname = 'William' and in
>>>>>>>>>>>> traverse(0,-1,'in') ( givenname = 'George' )
>>>>>>>>>>>>                                                     ^
>>>>>>>>>>>> Encountered "" at line 1, column 52.
>>>>>>>>>>>> Was expecting one of:
>>>>>>>>>>>>
>>>>>>>>>>>> DB name="familytree"
>>>>>>>>>>>>
>>>>>>>>>>>> I hope it hasn't been removed.  We use both the TRAVERSE FROM
>>>>>>>>>>>> and
>>>>>>>>>>>> SELECT FROM WHERE traverse() extensively.
>>>>>>>>>>>>
>>>>>>>>>>>> TRAVERSE FROM is wonderful for gathering a group of records
>>>>>>>>>>>> from a given point.
>>>>>>>>>>>> The traverse() operator is wonderful for queries like the above
>>>>>>>>>>>> (return all people named William
>>>>>>>>>>>> descended from a person named George.
>>>>>>>>>>>>
>>>>>>>>>>>> We really need both.
>>>>>>>>>>>>
>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>
>>>>>>>>>>>> Steven Tomer
>>>>>>>>>>>>
>>>>>>>>>>>> --
>>>>>>>>>>>>
>>>>>>>>>>>> ---
>>>>>>>>>>>> You received this message because you are subscribed to the
>>>>>>>>>>>> Google Groups "OrientDB" group.
>>>>>>>>>>>> To unsubscribe from this group and stop receiving emails from
>>>>>>>>>>>> it, send an email to orient-database+unsubscribe@
>>>>>>>>>>>> googlegroups.com.
>>>>>>>>>>>> For more options, visit https://groups.google.com/d/optout.
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> --
>>>>>>>>>>>
>>>>>>>>>>> ---
>>>>>>>>>>> You received this message because you are subscribed to a topic
>>>>>>>>>>> in the Google Groups "OrientDB" group.
>>>>>>>>>>> To unsubscribe from this topic, visit
>>>>>>>>>>> https://groups.google.com/d/topic/orient-database/ic5UE-
>>>>>>>>>>> kyi-w/unsubscribe.
>>>>>>>>>>> To unsubscribe from this group and all its topics, send an email
>>>>>>>>>>> to orient-database+unsubscr...@googlegroups.com.
>>>>>>>>>>> For more options, visit https://groups.google.com/d/optout.
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> --
>>>>>>>>>>
>>>>>>>>>> ---
>>>>>>>>>> You received this message because you are subscribed to the
>>>>>>>>>> Google Groups "OrientDB" group.
>>>>>>>>>> To unsubscribe from this group and stop receiving emails from it,
>>>>>>>>>> send an email to orient-database+unsubscr...@googlegroups.com.
>>>>>>>>>> For more options, visit https://groups.google.com/d/optout.
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>> --
>>>>>>>>>
>>>>>>>>> ---
>>>>>>>>> You received this message because you are subscribed to a topic in
>>>>>>>>> the Google Groups "OrientDB" group.
>>>>>>>>> To unsubscribe from this topic, visit https://groups.google.com/d/
>>>>>>>>> topic/orient-database/ic5UE-kyi-w/unsubscribe.
>>>>>>>>> To unsubscribe from this group and all its topics, send an email
>>>>>>>>> to orient-database+unsubscr...@googlegroups.com.
>>>>>>>>> For more options, visit https://groups.google.com/d/optout.
>>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>>
>>>>>>>> ---
>>>>>>>> You received this message because you are subscribed to the Google
>>>>>>>> Groups "OrientDB" group.
>>>>>>>> To unsubscribe from this group and stop receiving emails from it,
>>>>>>>> send an email to orient-database+unsubscr...@googlegroups.com.
>>>>>>>> For more options, visit https://groups.google.com/d/optout.
>>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>>
>>>>>>> ---
>>>>>>> You received this message because you are subscribed to a topic in
>>>>>>> the Google Groups "OrientDB" group.
>>>>>>> To unsubscribe from this topic, visit https://groups.google.com/d/
>>>>>>> topic/orient-database/ic5UE-kyi-w/unsubscribe.
>>>>>>> To unsubscribe from this group and all its topics, send an email to
>>>>>>> orient-database+unsubscr...@googlegroups.com.
>>>>>>> For more options, visit https://groups.google.com/d/optout.
>>>>>>>
>>>>>>
>>>>>> --
>>>>>>
>>>>>> ---
>>>>>> You received this message because you are subscribed to the Google
>>>>>> Groups "OrientDB" group.
>>>>>> To unsubscribe from this group and stop receiving emails from it,
>>>>>> send an email to orient-database+unsubscr...@googlegroups.com.
>>>>>> For more options, visit https://groups.google.com/d/optout.
>>>>>>
>>>>>
>>>>> --
>>>>>
>>>>> ---
>>>>> You received this message because you are subscribed to a topic in the
>>>>> Google Groups "OrientDB" group.
>>>>> To unsubscribe from this topic, visit https://groups.google.com/d/
>>>>> topic/orient-database/ic5UE-kyi-w/unsubscribe.
>>>>> To unsubscribe from this group and all its topics, send an email to
>>>>> orient-database+unsubscr...@googlegroups.com.
>>>>> For more options, visit https://groups.google.com/d/optout.
>>>>>
>>>>
>>>> --
>>>
>>> ---
>>> You received this message because you are subscribed to the Google
>>> Groups "OrientDB" group.
>>> To unsubscribe from this group and stop receiving emails from it, send
>>> an email to orient-database+unsubscr...@googlegroups.com.
>>> For more options, visit https://groups.google.com/d/optout.
>>>
>>
>> --
>>
>> ---
>> You received this message because you are subscribed to a topic in the
>> Google Groups "OrientDB" group.
>> To unsubscribe from this topic, visit https://groups.google.com/d/
>> topic/orient-database/ic5UE-kyi-w/unsubscribe.
>> To unsubscribe from this group and all its topics, send an email to
>> orient-database+unsubscr...@googlegroups.com.
>> For more options, visit https://groups.google.com/d/optout.
>>
>

-- 

--- 
You received this message because you are subscribed to the Google Groups 
"OrientDB" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to orient-database+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to