Use of ANTLR instead of CarbonSqlParser

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Use of ANTLR instead of CarbonSqlParser

anuragknoldus
Hi,

We are using CarbonSqlParser for parsing the queries but we can use ANTLR
for the same.

Is there any specific reason for using CarbonSqlParser because as per my
concern we could better handle with ANTLR.


--
*Thanks&regards*


*Anurag Srivastava**Software Consultant*
*Knoldus Software LLP*

*India - US - Canada*
* Twitter <http://www.twitter.com/anuragknoldus> | FB
<http://www.facebook.com/anuragsrivastava.06> | LinkedIn
<https://in.linkedin.com/pub/anurag-srivastava/5a/b6/441>*
Reply | Threaded
Open this post in threaded view
|

Re: Use of ANTLR instead of CarbonSqlParser

Jacky Li
Hi,

It is because CarbonData currently is integrated with Spark 1.5/1.6 and CarbonContext is based on HiveContext, so it is based on the Hive parser in HiveContext. But you are right, there is no design limitation about this, Carbon can switch to use ANTLR. I see in Spark 2.0, spark is using ANTLR as well. So maybe it is a good time to switch the parser when doing integration with Spark 2.0.
What is your idea?

Regards,
Jacky


> 在 2016年11月4日,下午3:47,Anurag Srivastava <[hidden email]> 写道:
>
> Hi,
>
> We are using CarbonSqlParser for parsing the queries but we can use ANTLR
> for the same.
>
> Is there any specific reason for using CarbonSqlParser because as per my
> concern we could better handle with ANTLR.
>
>
> --
> *Thanks&regards*
>
>
> *Anurag Srivastava**Software Consultant*
> *Knoldus Software LLP*
>
> *India - US - Canada*
> * Twitter <http://www.twitter.com/anuragknoldus> | FB
> <http://www.facebook.com/anuragsrivastava.06> | LinkedIn
> <https://in.linkedin.com/pub/anurag-srivastava/5a/b6/441>*

Reply | Threaded
Open this post in threaded view
|

Re: Use of ANTLR instead of CarbonSqlParser

anuragknoldus
Hi,

We can use ANTLR with Spark1.5/1.6. There is no compatibility issue. We can
replace the carbon parser with the ANTLR even when CarbonData is integrated
with Spark 1.5/1.6.  And when we are integrating with Spark 2.0, not much
changes will be required in the ANTLR parser.

But, if integration with Spark 2.0 is planned in near future, we can switch
to ANTLR parser at that time as well.

On Mon, Nov 7, 2016 at 6:59 AM, Jacky Li <[hidden email]> wrote:

> Hi,
>
> It is because CarbonData currently is integrated with Spark 1.5/1.6 and
> CarbonContext is based on HiveContext, so it is based on the Hive parser in
> HiveContext. But you are right, there is no design limitation about this,
> Carbon can switch to use ANTLR. I see in Spark 2.0, spark is using ANTLR as
> well. So maybe it is a good time to switch the parser when doing
> integration with Spark 2.0.
> What is your idea?
>
> Regards,
> Jacky
>
>
> > 在 2016年11月4日,下午3:47,Anurag Srivastava <[hidden email]> 写道:
> >
> > Hi,
> >
> > We are using CarbonSqlParser for parsing the queries but we can use ANTLR
> > for the same.
> >
> > Is there any specific reason for using CarbonSqlParser because as per my
> > concern we could better handle with ANTLR.
> >
> >
> > --
> > *Thanks&regards*
> >
> >
> > *Anurag Srivastava**Software Consultant*
> > *Knoldus Software LLP*
> >
> > *India - US - Canada*
> > * Twitter <http://www.twitter.com/anuragknoldus> | FB
> > <http://www.facebook.com/anuragsrivastava.06> | LinkedIn
> > <https://in.linkedin.com/pub/anurag-srivastava/5a/b6/441>*
>
>


--
*Thanks&regards*


*Anurag Srivastava**Software Consultant*
*Knoldus Software LLP*

*India - US - Canada*
* Twitter <http://www.twitter.com/anuragknoldus> | FB
<http://www.facebook.com/anuragsrivastava.06> | LinkedIn
<https://in.linkedin.com/pub/anurag-srivastava/5a/b6/441>*
Reply | Threaded
Open this post in threaded view
|

Re: Use of ANTLR instead of CarbonSqlParser

Jacky Li
Hi,

Yes, ANTLR parser is anyway needed, and yes we have planned to support spark 2 by the end of this year, so feel free to create a JIRA ticket for the parser

Regards,
Jacky

> 在 2016年11月7日,下午3:59,Anurag Srivastava <[hidden email]> 写道:
>
> Hi,
>
> We can use ANTLR with Spark1.5/1.6. There is no compatibility issue. We can
> replace the carbon parser with the ANTLR even when CarbonData is integrated
> with Spark 1.5/1.6.  And when we are integrating with Spark 2.0, not much
> changes will be required in the ANTLR parser.
>
> But, if integration with Spark 2.0 is planned in near future, we can switch
> to ANTLR parser at that time as well.
>
> On Mon, Nov 7, 2016 at 6:59 AM, Jacky Li <[hidden email]> wrote:
>
>> Hi,
>>
>> It is because CarbonData currently is integrated with Spark 1.5/1.6 and
>> CarbonContext is based on HiveContext, so it is based on the Hive parser in
>> HiveContext. But you are right, there is no design limitation about this,
>> Carbon can switch to use ANTLR. I see in Spark 2.0, spark is using ANTLR as
>> well. So maybe it is a good time to switch the parser when doing
>> integration with Spark 2.0.
>> What is your idea?
>>
>> Regards,
>> Jacky
>>
>>
>>> 在 2016年11月4日,下午3:47,Anurag Srivastava <[hidden email]> 写道:
>>>
>>> Hi,
>>>
>>> We are using CarbonSqlParser for parsing the queries but we can use ANTLR
>>> for the same.
>>>
>>> Is there any specific reason for using CarbonSqlParser because as per my
>>> concern we could better handle with ANTLR.
>>>
>>>
>>> --
>>> *Thanks&regards*
>>>
>>>
>>> *Anurag Srivastava**Software Consultant*
>>> *Knoldus Software LLP*
>>>
>>> *India - US - Canada*
>>> * Twitter <http://www.twitter.com/anuragknoldus> | FB
>>> <http://www.facebook.com/anuragsrivastava.06> | LinkedIn
>>> <https://in.linkedin.com/pub/anurag-srivastava/5a/b6/441>*
>>
>>
>
>
> --
> *Thanks&regards*
>
>
> *Anurag Srivastava**Software Consultant*
> *Knoldus Software LLP*
>
> *India - US - Canada*
> * Twitter <http://www.twitter.com/anuragknoldus> | FB
> <http://www.facebook.com/anuragsrivastava.06> | LinkedIn
> <https://in.linkedin.com/pub/anurag-srivastava/5a/b6/441>*