Support for Datastax java driver

[mevivs]

Analyze and see if this is feasible to provide Datastax java driver support.

[pires]

+1

[kadiyams]

Hello, I have recently worked on the Datastax Java driver. Can I work on this enhancement.

[mevivs]

@kadiyams
👍

Please let us know, if case we need to discuss something during implementation.

-Vivek

[kadiyams]

Thanks Vivek.
Using the DStax Driver we dont need to worry about the following things since it would automatically be handled at the driver level:
connection pooling
node discovery
automatic failover
load balancing
Coupled with it once we use the CQLTranslator in Thrift client we could have the clean client code.

[kadiyams]

Vivek,
I have gotten started with this and making good progress.
Quick comment/question: Datastax driver exclusively supports only CQL3.0 as defined by their git page "A Java client driver for Apache Cassandra. This driver works exclusively with the Cassandra Query Language version 3 (CQL3) and Cassandra's binary protocol."

1. Shall I make sure that this is checked while loading the configuration
   If there are any super columns or CQL2.0 stuff in the column families etc then should I just throw a not supported error for the Datastax client. That way we are putting it out there upfront and then then entire implementation should not worry about the CQL3.0 enabled check.
2. Untill the above question is answered, I will move forward and throw an "Unsupported: Only CQL3.0 is supported by Datastax client" error and bubble it up.

[mevivs]

1. Shall I make sure that this is checked while loading the configuration
If there are any super columns or CQL2.0 stuff in the column families etc then should I just throw a not supported error for the Datastax client. That way we are putting it out there upfront and then then entire implementation should not worry about the CQL3.0 enabled check.

Looks fine to me.

Thanks,
-Vivek

[mevivs]

@kadiyams

Any update on this?

-Vivek

[jfgosselin]

+1

I’m also interested into this mainly for the new feature like automatic paging (http://www.datastax.com/dev/blog/client-side-improvements-in-cassandra-2-0) which is only available with the CQL java driver.

Any update ?

Thanks

[mevivs]

Thanks for the link, will have a look. BTW, pagination is supported by Kundera over Cassandra in both ways(Thrift and CQL3).

https://github.com/impetus-opensource/Kundera/blob/trunk/src/kundera-cassandra/src/test/java/com/impetus/kundera/query/ResultIteratorTest.java

-Vivek

[jfgosselin]

Thanks Vivek , I'll take a look .

Based on this post from Datastax I thought it was only available via the CQL Native Protocol:
http://stackoverflow.com/questions/20394233/does-cql3-client-support-native-protocol

[kadiyams]

Vivek,
I lost the traction on this, since my job (changed jobs recently) has gotten hectic and am struggling with the dead lines. it would be another month before I could touch this. So please if anyone else is proceeding with continue. I will loop back in here after a month. Sorry for the delay in providing this feature.

[jfgosselin]

Hi Vivek,

Are you adding support for the Datastax java driver ?

JF

[mevivs]

Yes. I have started work on it

-Vivek

[ganeshkannan]

Hi Vivek,
Adding support for Datastax driver will be really helpful. I have been looking for something like Kundera for Cassandra for a while, and I think support for Datastax driver needs to be provided so enterprises like ours can easily start adopting Kundera.

Thank you.

[ganeshkannan]

Hi Vivek,
When do you think tentatively this issue will be closed?
Thanks

[mevivs]

Team is working on it and first version of Data stax java driver support is planned with 2.11 release

-Vivek

[trajar]

For those interested, I started a project 'normandra' (https://github.com/trajar/normandra) which provides a kind of 'JPA-light' API for cassandra via the DataStax 2.x driver. It started out of a need to have simple JPA operations (but not the full stack) via the DataStax library, which at the time wasn't available with Kundera. It reads JPA annotations, creates CQL3 schemas, and supports save/get operations. I recently added one-to-one and one-to-many relationships, id generators (using counters), and embedded ids. I am in the process of adding many-to-one relationships and simple queries. It could be a good starting point for Kundera team for adding DataStax support.

[mevivs]

Update on this
Added first cut support. Still working to add more junits and functionalities.

How to use?
Have a look at https://github.com/impetus-opensource/Kundera/blob/trunk/src/kundera-cassandra/cassandra-ds-driver/src/test/java/com/impetus/kundera/client/cassandra/crud/PersonEntityTest.java

Wiki page will be published with 2.11 release.

-Vivek

[mevivs]

Support added for most of features as:

1. Pagination
2. Id auto generation
3. Association
4. Composite key
5. Batch processing

a3d0043

-Vivek

[brunocrt]

I'm testing the pagination on a table with composite partition key and I'm getting just one row (1) as result set. I'm setting the fetch size for 500 but the iterator just return 1 item. There is not exception/erro on logs! I don't understand why! the table has thousands of rows already inserted on it.

String queryString= "SELECT e FROM OrderEvent e WHERE e.primaryKey.partitionKey.location='US' AND e.primaryKey.partitionKey.year=2015"; com.impetus.kundera.query.Query query = (com.impetus.kundera.query.Query) em.createQuery(queryString, entityClass); query.setFetchSize(500); Iterator<OrderEvent> rows = query.iterate(); while(rows.hasNext()) { resultList.add( rows.next() ); }

I'm using kundera-cassandra-ds-driver version 3.4 with cassandra-driver-core (com.datastax.casssandra) version 2.1.10.3

In my driver configuration I'm using the DSClientFactory and CQL version 3.0.0

[brunocrt]

It seems that the issue is related to the composite partition key, because I tried a similar query with other table without composite partition key and it worked perfectly

[brunocrt]

I captured more details after removing the WHERE clause from my query and an exception occured! The query generated by the Kundera driver seems to be limiting the results using LIMIT 1. That could be causing the issue.

11:10:11.556 [main] ERROR c.i.k.c.c.d.DSClient - Error while executing query SELECT * FROM "order_event" WHERE token("partitionKey") > token(OrderPartitionKey [year=2016, input_location=US]) LIMIT 1 ALLOW FILTERING.

Caused by: com.impetus.kundera.KunderaException: com.datastax.driver.core.exceptions.SyntaxError: line 1:90 no viable alternative at input '[' (... token("partitionKey") > token([OrderPartitionKey] [...)
at com.impetus.kundera.client.cassandra.dsdriver.DSClient.execute(DSClient.java:724)
at com.impetus.kundera.client.cassandra.dsdriver.DSClient.executeQuery(DSClient.java:526)
at com.impetus.client.cassandra.query.ResultIterator.populateEntities(ResultIterator.java:280)
at com.impetus.client.cassandra.query.ResultIterator.onCheckRelation(ResultIterator.java:226)