[jira] [Created] (IGNITE-6769) SQL: prototype for hand-made parser

[Anton Vinogradov (Jira)]

Vladimir Ozerov created IGNITE-6769:
---------------------------------------

             Summary: SQL: prototype for hand-made parser
                 Key: IGNITE-6769
                 URL: https://issues.apache.org/jira/browse/IGNITE-6769
             Project: Ignite
          Issue Type: Task
      Security Level: Public (Viewable by anyone)
          Components: sql
            Reporter: Vladimir Ozerov
            Assignee: Vladimir Ozerov
             Fix For: 2.4






--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

[Ray]

Hi Vladimir,

What's the purpose of this ticket?

Does it mean in the future Ignite will have its own SQL parser instead of
H2's parser Ignite is using now?



--
Sent from: http://apache-ignite-developers.2346864.n4.nabble.com/

[Vladimir Ozerov]

Hi Ray,

Yes, we are going to gradually move away from H2. This will be long
process, so first we create execution engine for simple queries, then for
more complex, etc.. Also it is very likely that our current distributed
execution engine and co-location model will be reworked somehow.

On Mon, Oct 30, 2017 at 9:51 AM, Ray <[hidden email]> wrote:

> Hi Vladimir,
>
> What's the purpose of this ticket?
>
> Does it mean in the future Ignite will have its own SQL parser instead of
> H2's parser Ignite is using now?
>
>
>
> --
> Sent from: http://apache-ignite-developers.2346864.n4.nabble.com/
>

[Ray]

Hi Vladimir,

Thanks for the reply.

Is there any roadmap for this new execution engine?
For example, in version 2.4 the new engine will process select commands and
the rest commands will still be handled by H2.





--
Sent from: http://apache-ignite-developers.2346864.n4.nabble.com/

[Vladimir Ozerov]

Hi Ray,

Not yet, I think we will heave a roadmap by the end of this year. I hope
that by the time of 2.4 release we will have the following:
1) Improve data loading speed [1]. As a part of this activity there is a
chance that we will change our implementation of our indexes, so that they
no longer depend on H2.
2) Have our own parser capable of handling DDL requests [2], [3]
3) Have a raw prototype of very simple execution engine to reason about
possible performance improvements (no ticket yet).

Vladimir.

[1]
https://cwiki.apache.org/confluence/display/IGNITE/IEP-1%3A+Bulk+data+loading+performance+improvements
[2] https://issues.apache.org/jira/browse/IGNITE-6769
[3] https://issues.apache.org/jira/browse/IGNITE-6276

On Mon, Oct 30, 2017 at 10:59 AM, Ray <[hidden email]> wrote:

> Hi Vladimir,
>
> Thanks for the reply.
>
> Is there any roadmap for this new execution engine?
> For example, in version 2.4 the new engine will process select commands and
> the rest commands will still be handled by H2.
>
>
>
>
>
> --
> Sent from: http://apache-ignite-developers.2346864.n4.nabble.com/
>