VDS using scratch table


We are using the scratch table to create tables with multi join & union.
We have deployed Dremio on Kubernetes and scratch table location is the default disk as provided in the default helm chart.

We created a VDS using this scratch table. To update the scratch table we drop and create the scratch table & drop and create the VDS reflection using SQL query.

Sometimes the reflection footprint of the VDS is 0.

Can using the scratch table in this way cause any issues.


@unni If reflection foot print is 0 then did the reflection not write any records, open the profile and see number of records the PARQUET_WRITER phase wrote or send us the profile