How to to do query push down by using TDV

[Sheikh Mohammad Zohaib]

I am using TDV to join two datasets from 2 platforms, I don't want TIBCO virtualization to process this data, Can I push down my query by using one of the databases and do the process on that database

[Sharath Chandrika Mummoju]

Hello,

Where to check if the view is pushed down to the database or not:

When you run any view in TDV, you can go check the TDV Studio -> Manager tab -> Requests panel (or) click on the 'Query execution' plan button on the view.

When we create any view in TDV and have a query in it --> execute it --> The query will be pushed down to the data source, gets executed and the data is retrieved.

But there may be some functions that are not pushed,it could be due to the fact that the data source might not support that function or TDVhas missed mapping that function, or they cannot be pushed further.

You can see the mapping in [TDV_Install_Dir]appsdlm\conf.capabilities file.

In the query plan, check what SQL you have in the FETCH node. The SQL in the FETCH node is what is pushed down to the data source.

TDV does not execute that SQL,that SQL is pushed to the data source and executes in the database and the result is returned to TDV.

Some functions or joins might not get pushed.. you would see them in the query plan. If these are not part of the FETCH node,it means these functions are applied or executed by TDV.

Sometimes, you would have no push reason messages in the query plan, explaining the reason why it could not be pushed further. Evaluate these messages and take appropriate action if needed.

Regards,

Chandrika

[Sheikh Mohammad Zohaib]

Chandrika thank you for your detailed reply. Issue is, we are reading from Hadoop system with where clause and in where clause we are reading numbers from Teradata system.

 

Our expectation is that Tibco will push that list of numbers to Hadoop system and fetch the result sets (assuming processing should be done within hadoop system). However in our case, tibco is reading from Teradata and reading from Hadoop and then doing processing on TDV level. 

 

Can you recommend some option, we have tried Data shipment option also but its not supported by TDV ,  also we have tried the cache as well.

 

Please help as we are stuck with this problem push down and our solution fails it doesn't work 

 

 

[Sharath Chandrika Mummoju]

Hello, this seems like a custom use case. Please open a ticket with the TDV Support team and they will determine if this is a supported use case or a PSG (Professional Service Group which is a paid service

) involvement would be needed to help further. Thank you.