Shuffle node

The shuffle node redistributes rows between workers to improve query parallelism.

The node displays in the query plan as S64 DA Hashed Shuffle. Shuffling is only supported when the parallel_leader_participation setting is disabled.

Settings

swarm64da.enable_shuffle (boolean)

Enables or disables the shuffle node. The default is on.

swarm64da.enable_shuffle_clause_minimization (boolean)

Enables or disables a shuffle node optimization, which reduces the clauses used to compute the redistribution. The default is on.

swarm64da.enable_shuffled_aggregate (boolean)

Enables or disables the use of the shuffle node to parallelize aggregates. The default is on.

swarm64da.enable_shuffled_distinct (boolean)

Enables or disables the use of the shuffle node to parallelize SELECT DISTINCT and SELECT DISTINCT ON statements. The default is on.