Query Pushdown in Snowflake

[Sam]

Hi,

I am wondering what documentation exists on Query Pushdown in Snowflake.

I noticed that a single function (monitonically_increasing_id()) prevented the entire query being pushed down to Snowflake during an ETL process.

Is Pushdown coming from the Spark JDBC connector OR is this from Databricks itself?

Is it possible to determine whether the query will be pushed down?

[Kaniz_Fatma]

Hi @ Sam ! My name is Kaniz, and I'm the technical moderator here. Great to meet you, and thanks for your question! Let's see if your peers on the community have an answer to your question first. Or else I will follow up shortly with a response.

[siddhathPanchal · Accepted solution]

Hi Sam,

The Spark Connector applies predicate and query pushdown by capturing and analyzing the Spark logical plans for SQL operations. When the data source is Snowflake, the operations are translated into a SQL query and then executed in Snowflake to improve performance.

However, because this translation requires almost a one-to-one translation of Spark SQL operators to Snowflake expressions, not all of Spark SQL operators can be pushed down. When pushdown fails, the connector falls back to a less-optimized execution plan. The unsupported operations are instead performed in Spark.

https://docs.snowflake.com/en/user-guide/spark-connector-use.html#pushdown