You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@calcite.apache.org by "Shiven Dvrk (Jira)" <ji...@apache.org> on 2022/03/22 00:04:00 UTC

[jira] [Created] (CALCITE-5054) Avatica Jdbc layer: Json vs ProtoBuf performance

Shiven Dvrk created CALCITE-5054:
------------------------------------

             Summary: Avatica Jdbc layer: Json vs ProtoBuf performance
                 Key: CALCITE-5054
                 URL: https://issues.apache.org/jira/browse/CALCITE-5054
             Project: Calcite
          Issue Type: Bug
          Components: avatica, jdbc-driver
            Reporter: Shiven Dvrk


We have run some benchmarks and also observed the similar behavior with our usecase testing, it looks like Protobuf serialization seems to be slower than JSON serialization at avatical layer.

 

We have used a simple query like this:

select Field1, Field2, Field3, Field4, Field5, Field6 from T3

Field1 to Field 3 are of SqlTypeName.BIGINT type

Field4 to Field6 are of SqlTypeName.DECIMAL type

We had case where we project more than 500K rows so we tried some variations on benchmarking but we see that protobuf is slower than json.

We are wondering why protobuf is slower? does anyone know if there is anything we might be missing?

 



--
This message was sent by Atlassian Jira
(v8.20.1#820001)