Review comment from Gian -
Add configs for to enable following behavior
- Ingestion-side, replace nulls with default values.
- Query-side, respect nulls if they are present in storage.
This allows the default configuration to be speedy (no nulls), while also allowing people to take advantage of more powerful null handling without too much work (they just need to change the storage-side config). Importantly, people don't need to set a cluster-wide config that changes the behavior for all datasources and all columns.
Review comment from Gian -
Add configs for to enable following behavior
This allows the default configuration to be speedy (no nulls), while also allowing people to take advantage of more powerful null handling without too much work (they just need to change the storage-side config). Importantly, people don't need to set a cluster-wide config that changes the behavior for all datasources and all columns.