Closed gangliao closed 2 years ago
YSB:
=== Stage 0 ===
CoalesceBatchesExec: target_batch_size=4096
RepartitionExec: partitioning=Hash([Column { name: "ad_id", index: 0 }], 16)
CoalesceBatchesExec: target_batch_size=4096
FilterExec: event_type@1 = view
RepartitionExec: partitioning=RoundRobinBatch(16)
MemoryExec: partitions=0, partition_sizes=[]
CoalesceBatchesExec: target_batch_size=4096
RepartitionExec: partitioning=Hash([Column { name: "c_ad_id", index: 0 }], 16)
RepartitionExec: partitioning=RoundRobinBatch(16)
MemoryExec: partitions=0, partition_sizes=[]
=== Stage 1 ===
CoalesceBatchesExec: target_batch_size=4096
RepartitionExec: partitioning=Hash([Column { name: "campaign_id", index: 0 }], 16)
HashAggregateExec: mode=Partial, gby=[campaign_id@3 as campaign_id], aggr=[COUNT(UInt8(1))]
CoalesceBatchesExec: target_batch_size=4096
HashJoinExec: mode=Partitioned, join_type=Inner, on=[(Column { name: "ad_id", index: 0 }, Column { name: "c_ad_id", index: 0 })]
MemoryExec: partitions=0, partition_sizes=[]
MemoryExec: partitions=0, partition_sizes=[]
=== Stage 2 ===
ProjectionExec: expr=[campaign_id@0 as campaign_id, COUNT(UInt8(1))@1 as COUNT(UInt8(1))]
HashAggregateExec: mode=FinalPartitioned, gby=[campaign_id@0 as campaign_id], aggr=[COUNT(UInt8(1))]
MemoryExec: partitions=0, partition_sizes=[]
check the next function type: lambda or group, then choose execute_partitioned or execute.
NEXMark Q4: