How Hana Integrates large size partitioned table data?

[former_member241108]

I have this question in my mind and thought of asking it , so i searched but couldn't find if any one has answered it.

Here is scenario : I have scale out HANA system with 3 hosts , one is master and other is slave. I have one large size table which is partitioned across these 3 hosts (partioning method is irrelevant here), my question is : if i query onto this 3 partitions, assume here that required portion of data is spread across all hosts, does HANA use map reduce type functions to integrate the data , if yes where integration happens? is it in master node?

[lbreddemann]

Well, obviously at least the result needs to be handed back to the client. How much of the stored base data actually gets transferred depends on the model, the query and the data.

[former_member241108]

Thanks for reply Lars , what about consolidation of data, will the transfer will happen from other nodes to node where connection is connected.

[lbreddemann]

Depending on the statement routing setup, the query will be mainly executed on the node to which the session is connected or where the majority of the data is located.