diff --git a/docs/how-to/query-options.md b/docs/how-to/query-options.md index e34c4e16d62..3fdaa55979d 100644 --- a/docs/how-to/query-options.md +++ b/docs/how-to/query-options.md @@ -7,9 +7,9 @@ Query options adjust selected behavior for a connection or request; they are not | `query.parallelism` | `query_parallelism`, `query.parallelism` | Integer `1`–`1024`; server default `0` means automatic CPU-based parallelism. This is not a resource-pool limit. | | `query.allow_query_fallback` | `query_fallback`, `allow_query_fallback`, `query.allow_query_fallback` (legacy `ALLOW_QUERY_FALLBACK` syntax / `query_fallback` hint also work) | `true` or `false` | | `query.enable_remote_dynamic_filter_pushdown` | canonical name | `true` or `false` | -| `datafusion.optimizer.*` | canonical name | Boolean; only `repartition_joins`, `repartition_aggregations`, `repartition_sorts`, `repartition_windows`, `enable_round_robin_repartition`, `prefer_existing_sort`, `prefer_hash_join`, `join_reordering`, `enable_topk_aggregation`, and `enable_dynamic_filter_pushdown` are supported. | +| `datafusion.optimizer.*` | canonical name | Keys, types, defaults, and validation come from the linked DataFusion version (for example, numeric `max_passes` and Boolean `filter_null_join_keys`). | -Use `SET query_parallelism = 4` or `SET query.parallelism TO 4`; `SHOW VARIABLES query.parallelism` works in MySQL and PostgreSQL, which also accepts `SHOW query.parallelism`. `SHOW VARIABLES