Slow read using Snowflake connector in Databricks.

[peta]

Hi,

I am trying to read table from Snowflake with Databricks native Snowflake jdbc connector. It is going well for small amount of data (100 rows), but if I am adding more (even just 1000 rows) the query does not finish. I was checking if the query finishes in Snowflake part and is takes just few miliseconds. Could you please advice where could be the problem?

[liquido]

Hi,

Have your issue got resolved, I'm facing the similar issue. If your issue got resolved can you please post the steps that you followed to resolve your issue?

Thanks

[peta]

Hi @Retired_mod ,

Thank you for your reply.

1. Network should not be a problem, the communication is between 2 Azure subscriptions, but I am checking it with our Operation team just to be sure.

2.We are using the smallest Snowflake warehouse (XS) and it seems working ok, all the queries I am sending there through JDBC ends with Success on Snowlake side but in Databricks we do not get the data. Futhermore I am running there really small queries, do you think that for the communication would be bigger warehouse better?

3. We are now using snowflake-jdbc 3.13.29 which should be the latest one. Which connection parameters do you thing can be tuned for better performace?

4. How can I use parquet files instead of csv through JDBC?

Thanks.

Petra 

[sri123]

Hi @peta 

Have your issue got resolved, I'm facing the similar issue. If your issue got resolved can you please post the steps that you followed to resolve your issue?

Thanks & Regards

Sri