databricks-sql-python icon indicating copy to clipboard operation
databricks-sql-python copied to clipboard

Original thrift file

Open chulucninh09 opened this issue 1 year ago • 3 comments

Hi, there is a gap between Hive server thrift file and the generated Python code. Where can I get the original thrift file of Python code?

chulucninh09 avatar Jun 09 '24 17:06 chulucninh09

Hi @chulucninh09! Can you please clarify what you mean? This library's Python code is up to date with thrift file used in Databricks

kravets-levko avatar Jun 09 '24 19:06 kravets-levko

I want to build other binding, but I cannot do that because I don't have original thrift file. When reading around, I got a hint that Databricks JDBC driver is Spark JDBC driver. And Spark JDBC is based on Hive Server 2. But when reading the detailed generated client code, it is not match with Hive Server 2 thrift file.

chulucninh09 avatar Jun 09 '24 22:06 chulucninh09

FYI Databricks have not publicly shared the original .thrift file that is consumed by Databricks compute. You're correct that the open source HS2 .thrift file is different from what Databricks use in production. If you're trying to develop a new thrift binding I'd recommeond speaking with Databricks customer support to get more information. It's unlikely that the thrift definition would be published in the Python connector repository (where you opened this issue).

susodapop avatar Jun 12 '24 16:06 susodapop