http://git-wip-us.apache.org/repos/asf/hive/blob/83d1fd23/ql/src/test/results/clientpositive/perf/tez/constraints/query23.q.out ---------------------------------------------------------------------- diff --git a/ql/src/test/results/clientpositive/perf/tez/constraints/query23.q.out b/ql/src/test/results/clientpositive/perf/tez/constraints/query23.q.out index 7678363..f848093 100644 --- a/ql/src/test/results/clientpositive/perf/tez/constraints/query23.q.out +++ b/ql/src/test/results/clientpositive/perf/tez/constraints/query23.q.out @@ -1,7 +1,7 @@ -Warning: Shuffle Join MERGEJOIN[437][tables = [$hdt$_1, $hdt$_2]] in Stage 'Reducer 19' is a cross product -Warning: Shuffle Join MERGEJOIN[438][tables = [$hdt$_1, $hdt$_2, $hdt$_0]] in Stage 'Reducer 20' is a cross product -Warning: Shuffle Join MERGEJOIN[440][tables = [$hdt$_2, $hdt$_3]] in Stage 'Reducer 23' is a cross product -Warning: Shuffle Join MERGEJOIN[441][tables = [$hdt$_2, $hdt$_3, $hdt$_1]] in Stage 'Reducer 24' is a cross product +Warning: Shuffle Join MERGEJOIN[443][tables = [$hdt$_1, $hdt$_2]] in Stage 'Reducer 19' is a cross product +Warning: Shuffle Join MERGEJOIN[444][tables = [$hdt$_1, $hdt$_2, $hdt$_0]] in Stage 'Reducer 20' is a cross product +Warning: Shuffle Join MERGEJOIN[446][tables = [$hdt$_1, $hdt$_2]] in Stage 'Reducer 23' is a cross product +Warning: Shuffle Join MERGEJOIN[447][tables = [$hdt$_1, $hdt$_2, $hdt$_0]] in Stage 'Reducer 24' is a cross product PREHOOK: query: explain with frequent_ss_items as (select substr(i_item_desc,1,30) itemdesc,i_item_sk item_sk,d_date solddate,count(*) cnt @@ -123,13 +123,13 @@ Plan optimized by CBO. Vertex dependency in root stage Map 1 <- Reducer 9 (BROADCAST_EDGE) Map 15 <- Reducer 27 (BROADCAST_EDGE) -Map 33 <- Reducer 7 (BROADCAST_EDGE) -Map 35 <- Reducer 32 (BROADCAST_EDGE) -Map 37 <- Reducer 14 (BROADCAST_EDGE) -Map 38 <- Reducer 13 (BROADCAST_EDGE) -Reducer 10 <- Map 37 (SIMPLE_EDGE), Map 8 (SIMPLE_EDGE) +Map 28 <- Reducer 7 (BROADCAST_EDGE) +Map 30 <- Reducer 36 (BROADCAST_EDGE) +Map 38 <- Reducer 14 (BROADCAST_EDGE) +Map 39 <- Reducer 13 (BROADCAST_EDGE) +Reducer 10 <- Map 38 (SIMPLE_EDGE), Map 8 (SIMPLE_EDGE) Reducer 11 <- Reducer 10 (SIMPLE_EDGE), Reducer 24 (SIMPLE_EDGE) -Reducer 12 <- Reducer 11 (SIMPLE_EDGE), Reducer 31 (ONE_TO_ONE_EDGE), Union 5 (CONTAINS) +Reducer 12 <- Reducer 11 (SIMPLE_EDGE), Reducer 34 (SIMPLE_EDGE), Union 5 (CONTAINS) Reducer 13 <- Reducer 10 (CUSTOM_SIMPLE_EDGE) Reducer 14 <- Map 8 (CUSTOM_SIMPLE_EDGE) Reducer 16 <- Map 15 (SIMPLE_EDGE), Map 26 (SIMPLE_EDGE) @@ -137,22 +137,22 @@ Reducer 17 <- Reducer 16 (SIMPLE_EDGE) Reducer 18 <- Reducer 17 (CUSTOM_SIMPLE_EDGE) Reducer 19 <- Reducer 18 (CUSTOM_SIMPLE_EDGE), Reducer 21 (CUSTOM_SIMPLE_EDGE) Reducer 2 <- Map 1 (SIMPLE_EDGE), Map 8 (SIMPLE_EDGE) -Reducer 20 <- Reducer 19 (CUSTOM_SIMPLE_EDGE), Reducer 34 (CUSTOM_SIMPLE_EDGE) +Reducer 20 <- Reducer 19 (CUSTOM_SIMPLE_EDGE), Reducer 29 (CUSTOM_SIMPLE_EDGE) Reducer 21 <- Reducer 17 (CUSTOM_SIMPLE_EDGE) Reducer 22 <- Reducer 17 (CUSTOM_SIMPLE_EDGE) Reducer 23 <- Reducer 22 (CUSTOM_SIMPLE_EDGE), Reducer 25 (CUSTOM_SIMPLE_EDGE) -Reducer 24 <- Reducer 23 (CUSTOM_SIMPLE_EDGE), Reducer 39 (CUSTOM_SIMPLE_EDGE) +Reducer 24 <- Reducer 23 (CUSTOM_SIMPLE_EDGE), Reducer 40 (CUSTOM_SIMPLE_EDGE) Reducer 25 <- Reducer 17 (CUSTOM_SIMPLE_EDGE) Reducer 27 <- Map 26 (CUSTOM_SIMPLE_EDGE) -Reducer 28 <- Map 26 (SIMPLE_EDGE), Map 35 (SIMPLE_EDGE) -Reducer 29 <- Map 36 (SIMPLE_EDGE), Reducer 28 (SIMPLE_EDGE) +Reducer 29 <- Map 28 (SIMPLE_EDGE) Reducer 3 <- Reducer 2 (SIMPLE_EDGE), Reducer 20 (SIMPLE_EDGE) -Reducer 30 <- Reducer 29 (SIMPLE_EDGE) -Reducer 31 <- Reducer 29 (SIMPLE_EDGE) -Reducer 32 <- Map 26 (CUSTOM_SIMPLE_EDGE) -Reducer 34 <- Map 33 (SIMPLE_EDGE) -Reducer 39 <- Map 38 (SIMPLE_EDGE) -Reducer 4 <- Reducer 3 (SIMPLE_EDGE), Reducer 30 (ONE_TO_ONE_EDGE), Union 5 (CONTAINS) +Reducer 31 <- Map 30 (SIMPLE_EDGE), Map 35 (SIMPLE_EDGE) +Reducer 32 <- Map 37 (SIMPLE_EDGE), Reducer 31 (SIMPLE_EDGE) +Reducer 33 <- Reducer 32 (SIMPLE_EDGE) +Reducer 34 <- Reducer 32 (SIMPLE_EDGE) +Reducer 36 <- Map 35 (CUSTOM_SIMPLE_EDGE) +Reducer 4 <- Reducer 3 (SIMPLE_EDGE), Reducer 33 (SIMPLE_EDGE), Union 5 (CONTAINS) +Reducer 40 <- Map 39 (SIMPLE_EDGE) Reducer 6 <- Union 5 (CUSTOM_SIMPLE_EDGE) Reducer 7 <- Reducer 2 (CUSTOM_SIMPLE_EDGE) Reducer 9 <- Map 8 (CUSTOM_SIMPLE_EDGE) @@ -162,335 +162,348 @@ Stage-0 limit:100 Stage-1 Reducer 6 vectorized - File Output Operator [FS_527] - Limit [LIM_526] (rows=1 width=112) + File Output Operator [FS_537] + Limit [LIM_536] (rows=1 width=112) Number of rows:100 - Group By Operator [GBY_525] (rows=1 width=112) + Group By Operator [GBY_535] (rows=1 width=112) Output:["_col0"],aggregations:["sum(VALUE._col0)"] <-Union 5 [CUSTOM_SIMPLE_EDGE] <-Reducer 12 [CONTAINS] - Reduce Output Operator [RS_454] - Group By Operator [GBY_453] (rows=1 width=112) + Reduce Output Operator [RS_460] + Group By Operator [GBY_459] (rows=1 width=112) Output:["_col0"],aggregations:["sum(_col0)"] - Select Operator [SEL_451] (rows=52 width=112) + Select Operator [SEL_457] (rows=155 width=112) Output:["_col0"] - Merge Join Operator [MERGEJOIN_450] (rows=52 width=2) - Conds:RS_192._col1=RS_552._col0(Inner),Output:["_col3","_col4"] + Merge Join Operator [MERGEJOIN_456] (rows=155 width=0) + Conds:RS_198._col1=RS_564._col0(Left Semi),Output:["_col3","_col4"] <-Reducer 11 [SIMPLE_EDGE] - SHUFFLE [RS_192] + SHUFFLE [RS_198] PartitionCols:_col1 - Merge Join Operator [MERGEJOIN_442] (rows=3941101 width=118) - Conds:RS_189._col2=RS_190._col0(Inner),Output:["_col1","_col3","_col4"] + Merge Join Operator [MERGEJOIN_448] (rows=3941101 width=118) + Conds:RS_193._col2=RS_194._col0(Inner),Output:["_col1","_col3","_col4"] <-Reducer 10 [SIMPLE_EDGE] - PARTITION_ONLY_SHUFFLE [RS_189] + PARTITION_ONLY_SHUFFLE [RS_193] PartitionCols:_col2 - Merge Join Operator [MERGEJOIN_432] (rows=3941102 width=122) - Conds:RS_532._col0=RS_459._col0(Inner),Output:["_col1","_col2","_col3","_col4"] + Merge Join Operator [MERGEJOIN_438] (rows=3941102 width=122) + Conds:RS_542._col0=RS_465._col0(Inner),Output:["_col1","_col2","_col3","_col4"] <-Map 8 [SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_459] + PARTITION_ONLY_SHUFFLE [RS_465] PartitionCols:_col0 - Select Operator [SEL_456] (rows=50 width=4) + Select Operator [SEL_462] (rows=50 width=4) Output:["_col0"] - Filter Operator [FIL_455] (rows=50 width=12) + Filter Operator [FIL_461] (rows=50 width=12) predicate:((d_moy = 1) and (d_year = 1999)) TableScan [TS_3] (rows=73049 width=12) default@date_dim,date_dim,Tbl:COMPLETE,Col:COMPLETE,Output:["d_date_sk","d_year","d_moy"] - <-Map 37 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_532] + <-Map 38 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_542] PartitionCols:_col0 - Select Operator [SEL_531] (rows=143930993 width=127) + Select Operator [SEL_541] (rows=143930993 width=127) Output:["_col0","_col1","_col2","_col3","_col4"] - Filter Operator [FIL_530] (rows=143930993 width=127) - predicate:((ws_sold_date_sk BETWEEN DynamicValue(RS_187_date_dim_d_date_sk_min) AND DynamicValue(RS_187_date_dim_d_date_sk_max) and in_bloom_filter(ws_sold_date_sk, DynamicValue(RS_187_date_dim_d_date_sk_bloom_filter))) and ws_bill_customer_sk is not null and ws_sold_date_sk is not null) - TableScan [TS_98] (rows=144002668 width=127) + Filter Operator [FIL_540] (rows=143930993 width=127) + predicate:((ws_sold_date_sk BETWEEN DynamicValue(RS_191_date_dim_d_date_sk_min) AND DynamicValue(RS_191_date_dim_d_date_sk_max) and in_bloom_filter(ws_sold_date_sk, DynamicValue(RS_191_date_dim_d_date_sk_bloom_filter))) and ws_bill_customer_sk is not null and ws_sold_date_sk is not null) + TableScan [TS_101] (rows=144002668 width=127) default@web_sales,web_sales,Tbl:COMPLETE,Col:COMPLETE,Output:["ws_sold_date_sk","ws_item_sk","ws_bill_customer_sk","ws_quantity","ws_list_price"] <-Reducer 14 [BROADCAST_EDGE] vectorized - BROADCAST [RS_529] - Group By Operator [GBY_528] (rows=1 width=12) + BROADCAST [RS_539] + Group By Operator [GBY_538] (rows=1 width=12) Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] <-Map 8 [CUSTOM_SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_464] - Group By Operator [GBY_462] (rows=1 width=12) + PARTITION_ONLY_SHUFFLE [RS_470] + Group By Operator [GBY_468] (rows=1 width=12) Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] - Select Operator [SEL_460] (rows=50 width=4) + Select Operator [SEL_466] (rows=50 width=4) Output:["_col0"] - Please refer to the previous Select Operator [SEL_456] + Please refer to the previous Select Operator [SEL_462] <-Reducer 24 [SIMPLE_EDGE] - SHUFFLE [RS_190] + SHUFFLE [RS_194] PartitionCols:_col0 - Select Operator [SEL_165] (rows=471875 width=3) + Select Operator [SEL_168] (rows=471875 width=3) Output:["_col0"] - Filter Operator [FIL_164] (rows=471875 width=227) + Filter Operator [FIL_167] (rows=471875 width=227) predicate:(_col3 > (0.95 * _col1)) - Merge Join Operator [MERGEJOIN_441] (rows=1415626 width=227) + Merge Join Operator [MERGEJOIN_447] (rows=1415626 width=227) Conds:(Inner),Output:["_col1","_col2","_col3"] <-Reducer 23 [CUSTOM_SIMPLE_EDGE] - PARTITION_ONLY_SHUFFLE [RS_161] - Merge Join Operator [MERGEJOIN_440] (rows=1 width=112) + PARTITION_ONLY_SHUFFLE [RS_164] + Merge Join Operator [MERGEJOIN_446] (rows=1 width=112) Conds:(Inner),Output:["_col1"] <-Reducer 22 [CUSTOM_SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_538] - Select Operator [SEL_537] (rows=1 width=8) - Filter Operator [FIL_536] (rows=1 width=8) + PARTITION_ONLY_SHUFFLE [RS_548] + Select Operator [SEL_547] (rows=1 width=8) + Filter Operator [FIL_546] (rows=1 width=8) predicate:(sq_count_check(_col0) <= 1) - Group By Operator [GBY_535] (rows=1 width=8) + Group By Operator [GBY_545] (rows=1 width=8) Output:["_col0"],aggregations:["count()"] - Select Operator [SEL_534] (rows=1 width=8) - Group By Operator [GBY_533] (rows=1 width=8) + Select Operator [SEL_544] (rows=1 width=8) + Group By Operator [GBY_543] (rows=1 width=8) Output:["_col0"],aggregations:["count(VALUE._col0)"] <-Reducer 17 [CUSTOM_SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_496] - Group By Operator [GBY_492] (rows=1 width=8) + PARTITION_ONLY_SHUFFLE [RS_498] + Group By Operator [GBY_494] (rows=1 width=8) Output:["_col0"],aggregations:["count(_col0)"] - Select Operator [SEL_488] (rows=50562 width=112) + Select Operator [SEL_490] (rows=50562 width=112) Output:["_col0"] - Group By Operator [GBY_485] (rows=50562 width=112) + Group By Operator [GBY_487] (rows=50562 width=112) Output:["_col0","_col1"],aggregations:["sum(VALUE._col0)"],keys:KEY._col0 <-Reducer 16 [SIMPLE_EDGE] SHUFFLE [RS_17] PartitionCols:_col0 Group By Operator [GBY_16] (rows=455058 width=112) Output:["_col0","_col1"],aggregations:["sum(_col2)"],keys:_col1 - Merge Join Operator [MERGEJOIN_428] (rows=18762463 width=112) - Conds:RS_484._col0=RS_472._col0(Inner),Output:["_col1","_col2"] + Merge Join Operator [MERGEJOIN_434] (rows=18762463 width=112) + Conds:RS_486._col0=RS_478._col0(Inner),Output:["_col1","_col2"] <-Map 26 [SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_472] + PARTITION_ONLY_SHUFFLE [RS_478] PartitionCols:_col0 - Select Operator [SEL_471] (rows=2609 width=4) + Select Operator [SEL_477] (rows=2609 width=4) Output:["_col0"] - Filter Operator [FIL_470] (rows=2609 width=8) + Filter Operator [FIL_476] (rows=2609 width=8) predicate:(d_year) IN (1999, 2000, 2001, 2002) TableScan [TS_9] (rows=73049 width=8) default@date_dim,date_dim,Tbl:COMPLETE,Col:COMPLETE,Output:["d_date_sk","d_year"] <-Map 15 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_484] + SHUFFLE [RS_486] PartitionCols:_col0 - Select Operator [SEL_483] (rows=525327388 width=119) + Select Operator [SEL_485] (rows=525327388 width=119) Output:["_col0","_col1","_col2"] - Filter Operator [FIL_482] (rows=525327388 width=118) + Filter Operator [FIL_484] (rows=525327388 width=118) predicate:((ss_sold_date_sk BETWEEN DynamicValue(RS_13_date_dim_d_date_sk_min) AND DynamicValue(RS_13_date_dim_d_date_sk_max) and in_bloom_filter(ss_sold_date_sk, DynamicValue(RS_13_date_dim_d_date_sk_bloom_filter))) and ss_customer_sk is not null and ss_sold_date_sk is not null) TableScan [TS_6] (rows=575995635 width=118) default@store_sales,store_sales,Tbl:COMPLETE,Col:COMPLETE,Output:["ss_sold_date_sk","ss_customer_sk","ss_quantity","ss_sales_price"] <-Reducer 27 [BROADCAST_EDGE] vectorized - BROADCAST [RS_481] - Group By Operator [GBY_480] (rows=1 width=12) + BROADCAST [RS_483] + Group By Operator [GBY_482] (rows=1 width=12) Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] <-Map 26 [CUSTOM_SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_478] - Group By Operator [GBY_476] (rows=1 width=12) + PARTITION_ONLY_SHUFFLE [RS_481] + Group By Operator [GBY_480] (rows=1 width=12) Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] - Select Operator [SEL_473] (rows=2609 width=4) + Select Operator [SEL_479] (rows=2609 width=4) Output:["_col0"] - Please refer to the previous Select Operator [SEL_471] + Please refer to the previous Select Operator [SEL_477] <-Reducer 25 [CUSTOM_SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_540] - Group By Operator [GBY_539] (rows=1 width=112) + PARTITION_ONLY_SHUFFLE [RS_550] + Group By Operator [GBY_549] (rows=1 width=112) Output:["_col0"],aggregations:["max(VALUE._col0)"] <-Reducer 17 [CUSTOM_SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_497] - Group By Operator [GBY_493] (rows=1 width=112) + PARTITION_ONLY_SHUFFLE [RS_499] + Group By Operator [GBY_495] (rows=1 width=112) Output:["_col0"],aggregations:["max(_col1)"] - Select Operator [SEL_489] (rows=50562 width=112) + Select Operator [SEL_491] (rows=50562 width=112) Output:["_col1"] - Please refer to the previous Group By Operator [GBY_485] - <-Reducer 39 [CUSTOM_SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_548] - Group By Operator [GBY_547] (rows=1415626 width=115) + Please refer to the previous Group By Operator [GBY_487] + <-Reducer 40 [CUSTOM_SIMPLE_EDGE] vectorized + PARTITION_ONLY_SHUFFLE [RS_558] + Group By Operator [GBY_557] (rows=1415626 width=115) Output:["_col0","_col1"],aggregations:["sum(VALUE._col0)"],keys:KEY._col0 - <-Map 38 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_546] + <-Map 39 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_556] PartitionCols:_col0 - Group By Operator [GBY_545] (rows=550080312 width=115) + Group By Operator [GBY_555] (rows=550080312 width=115) Output:["_col0","_col1"],aggregations:["sum(_col1)"],keys:_col0 - Select Operator [SEL_544] (rows=550080312 width=114) + Select Operator [SEL_554] (rows=550080312 width=114) Output:["_col0","_col1"] - Filter Operator [FIL_543] (rows=550080312 width=114) - predicate:((ss_customer_sk BETWEEN DynamicValue(RS_189_web_sales_ws_bill_customer_sk_min) AND DynamicValue(RS_189_web_sales_ws_bill_customer_sk_max) and in_bloom_filter(ss_customer_sk, DynamicValue(RS_189_web_sales_ws_bill_customer_sk_bloom_filter))) and ss_customer_sk is not null) - TableScan [TS_150] (rows=575995635 width=114) + Filter Operator [FIL_553] (rows=550080312 width=114) + predicate:((ss_customer_sk BETWEEN DynamicValue(RS_193_web_sales_ws_bill_customer_sk_min) AND DynamicValue(RS_193_web_sales_ws_bill_customer_sk_max) and in_bloom_filter(ss_customer_sk, DynamicValue(RS_193_web_sales_ws_bill_customer_sk_bloom_filter))) and ss_customer_sk is not null) + TableScan [TS_153] (rows=575995635 width=114) default@store_sales,store_sales,Tbl:COMPLETE,Col:COMPLETE,Output:["ss_customer_sk","ss_quantity","ss_sales_price"] <-Reducer 13 [BROADCAST_EDGE] vectorized - BROADCAST [RS_542] - Group By Operator [GBY_541] (rows=1 width=12) + BROADCAST [RS_552] + Group By Operator [GBY_551] (rows=1 width=12) Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] <-Reducer 10 [CUSTOM_SIMPLE_EDGE] - PARTITION_ONLY_SHUFFLE [RS_406] - Group By Operator [GBY_405] (rows=1 width=12) + PARTITION_ONLY_SHUFFLE [RS_412] + Group By Operator [GBY_411] (rows=1 width=12) Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] - Select Operator [SEL_404] (rows=3941102 width=7) + Select Operator [SEL_410] (rows=3941102 width=7) Output:["_col0"] - Please refer to the previous Merge Join Operator [MERGEJOIN_432] - <-Reducer 31 [ONE_TO_ONE_EDGE] vectorized - FORWARD [RS_552] + Please refer to the previous Merge Join Operator [MERGEJOIN_438] + <-Reducer 34 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_564] PartitionCols:_col0 - Select Operator [SEL_551] (rows=745 width=4) - Output:["_col0"] - Filter Operator [FIL_550] (rows=745 width=12) - predicate:(_col1 > 4L) - Group By Operator [GBY_549] (rows=2235 width=12) - Output:["_col0","_col1"],aggregations:["count(VALUE._col0)"],keys:KEY._col0 - <-Reducer 29 [SIMPLE_EDGE] - SHUFFLE [RS_182] - PartitionCols:_col0 - Group By Operator [GBY_83] (rows=2235 width=12) - Output:["_col0","_col1"],aggregations:["count()"],keys:_col4 - Merge Join Operator [MERGEJOIN_431] (rows=19646398 width=4) - Conds:RS_79._col1=RS_520._col0(Inner),Output:["_col4"] - <-Map 36 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_520] - PartitionCols:_col0 - Select Operator [SEL_519] (rows=462000 width=188) - Output:["_col0"] - TableScan [TS_74] (rows=462000 width=4) - default@item,item,Tbl:COMPLETE,Col:COMPLETE,Output:["i_item_sk"] - <-Reducer 28 [SIMPLE_EDGE] - SHUFFLE [RS_79] - PartitionCols:_col1 - Merge Join Operator [MERGEJOIN_430] (rows=19646398 width=4) - Conds:RS_518._col0=RS_474._col0(Inner),Output:["_col1"] - <-Map 26 [SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_474] - PartitionCols:_col0 - Please refer to the previous Select Operator [SEL_471] - <-Map 35 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_518] + Group By Operator [GBY_563] (rows=2235 width=4) + Output:["_col0"],keys:_col0 + Select Operator [SEL_562] (rows=6548799 width=4) + Output:["_col0"] + Filter Operator [FIL_561] (rows=6548799 width=290) + predicate:(_col3 > 4L) + Select Operator [SEL_560] (rows=19646398 width=290) + Output:["_col1","_col3"] + Group By Operator [GBY_559] (rows=19646398 width=290) + Output:["_col0","_col1","_col2","_col3"],aggregations:["count(VALUE._col0)"],keys:KEY._col0, KEY._col1, KEY._col2 + <-Reducer 32 [SIMPLE_EDGE] + SHUFFLE [RS_185] + PartitionCols:_col0, _col1, _col2 + Group By Operator [GBY_83] (rows=19646398 width=290) + Output:["_col0","_col1","_col2","_col3"],aggregations:["count()"],keys:_col4, _col3, _col5 + Merge Join Operator [MERGEJOIN_437] (rows=19646398 width=282) + Conds:RS_79._col1=RS_528._col0(Inner),Output:["_col3","_col4","_col5"] + <-Map 37 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_528] PartitionCols:_col0 - Select Operator [SEL_517] (rows=550076554 width=7) + Select Operator [SEL_527] (rows=462000 width=188) Output:["_col0","_col1"] - Filter Operator [FIL_516] (rows=550076554 width=7) - predicate:((ss_sold_date_sk BETWEEN DynamicValue(RS_77_date_dim_d_date_sk_min) AND DynamicValue(RS_77_date_dim_d_date_sk_max) and in_bloom_filter(ss_sold_date_sk, DynamicValue(RS_77_date_dim_d_date_sk_bloom_filter))) and ss_sold_date_sk is not null) - TableScan [TS_68] (rows=575995635 width=7) - default@store_sales,store_sales,Tbl:COMPLETE,Col:COMPLETE,Output:["ss_sold_date_sk","ss_item_sk"] - <-Reducer 32 [BROADCAST_EDGE] vectorized - BROADCAST [RS_515] - Group By Operator [GBY_514] (rows=1 width=12) - Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] - <-Map 26 [CUSTOM_SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_479] - Group By Operator [GBY_477] (rows=1 width=12) - Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] - Select Operator [SEL_475] (rows=2609 width=4) - Output:["_col0"] - Please refer to the previous Select Operator [SEL_471] + TableScan [TS_74] (rows=462000 width=188) + default@item,item,Tbl:COMPLETE,Col:COMPLETE,Output:["i_item_sk","i_item_desc"] + <-Reducer 31 [SIMPLE_EDGE] + SHUFFLE [RS_79] + PartitionCols:_col1 + Merge Join Operator [MERGEJOIN_436] (rows=19646398 width=98) + Conds:RS_526._col0=RS_518._col0(Inner),Output:["_col1","_col3"] + <-Map 35 [SIMPLE_EDGE] vectorized + PARTITION_ONLY_SHUFFLE [RS_518] + PartitionCols:_col0 + Select Operator [SEL_517] (rows=2609 width=98) + Output:["_col0","_col1"] + Filter Operator [FIL_516] (rows=2609 width=102) + predicate:(d_year) IN (1999, 2000, 2001, 2002) + TableScan [TS_71] (rows=73049 width=102) + default@date_dim,date_dim,Tbl:COMPLETE,Col:COMPLETE,Output:["d_date_sk","d_date","d_year"] + <-Map 30 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_526] + PartitionCols:_col0 + Select Operator [SEL_525] (rows=550076554 width=7) + Output:["_col0","_col1"] + Filter Operator [FIL_524] (rows=550076554 width=7) + predicate:((ss_sold_date_sk BETWEEN DynamicValue(RS_77_date_dim_d_date_sk_min) AND DynamicValue(RS_77_date_dim_d_date_sk_max) and in_bloom_filter(ss_sold_date_sk, DynamicValue(RS_77_date_dim_d_date_sk_bloom_filter))) and ss_sold_date_sk is not null) + TableScan [TS_68] (rows=575995635 width=7) + default@store_sales,store_sales,Tbl:COMPLETE,Col:COMPLETE,Output:["ss_sold_date_sk","ss_item_sk"] + <-Reducer 36 [BROADCAST_EDGE] vectorized + BROADCAST [RS_523] + Group By Operator [GBY_522] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] + <-Map 35 [CUSTOM_SIMPLE_EDGE] vectorized + PARTITION_ONLY_SHUFFLE [RS_521] + Group By Operator [GBY_520] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] + Select Operator [SEL_519] (rows=2609 width=4) + Output:["_col0"] + Please refer to the previous Select Operator [SEL_517] <-Reducer 4 [CONTAINS] - Reduce Output Operator [RS_449] - Group By Operator [GBY_448] (rows=1 width=112) + Reduce Output Operator [RS_455] + Group By Operator [GBY_454] (rows=1 width=112) Output:["_col0"],aggregations:["sum(_col0)"] - Select Operator [SEL_446] (rows=102 width=112) + Select Operator [SEL_452] (rows=304 width=112) Output:["_col0"] - Merge Join Operator [MERGEJOIN_445] (rows=102 width=1) - Conds:RS_94._col2=RS_524._col0(Inner),Output:["_col3","_col4"] + Merge Join Operator [MERGEJOIN_451] (rows=304 width=0) + Conds:RS_97._col2=RS_534._col0(Left Semi),Output:["_col3","_col4"] <-Reducer 3 [SIMPLE_EDGE] - SHUFFLE [RS_94] + SHUFFLE [RS_97] PartitionCols:_col2 - Merge Join Operator [MERGEJOIN_439] (rows=7751875 width=98) - Conds:RS_91._col1=RS_92._col0(Inner),Output:["_col2","_col3","_col4"] + Merge Join Operator [MERGEJOIN_445] (rows=7751875 width=98) + Conds:RS_92._col1=RS_93._col0(Inner),Output:["_col2","_col3","_col4"] <-Reducer 2 [SIMPLE_EDGE] - PARTITION_ONLY_SHUFFLE [RS_91] + PARTITION_ONLY_SHUFFLE [RS_92] PartitionCols:_col1 - Merge Join Operator [MERGEJOIN_427] (rows=7751875 width=101) - Conds:RS_469._col0=RS_457._col0(Inner),Output:["_col1","_col2","_col3","_col4"] + Merge Join Operator [MERGEJOIN_433] (rows=7751875 width=101) + Conds:RS_475._col0=RS_463._col0(Inner),Output:["_col1","_col2","_col3","_col4"] <-Map 8 [SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_457] + PARTITION_ONLY_SHUFFLE [RS_463] PartitionCols:_col0 - Please refer to the previous Select Operator [SEL_456] + Please refer to the previous Select Operator [SEL_462] <-Map 1 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_469] + SHUFFLE [RS_475] PartitionCols:_col0 - Select Operator [SEL_468] (rows=285117831 width=127) + Select Operator [SEL_474] (rows=285117831 width=127) Output:["_col0","_col1","_col2","_col3","_col4"] - Filter Operator [FIL_467] (rows=285117831 width=127) - predicate:((cs_sold_date_sk BETWEEN DynamicValue(RS_89_date_dim_d_date_sk_min) AND DynamicValue(RS_89_date_dim_d_date_sk_max) and in_bloom_filter(cs_sold_date_sk, DynamicValue(RS_89_date_dim_d_date_sk_bloom_filter))) and cs_bill_customer_sk is not null and cs_sold_date_sk is not null) + Filter Operator [FIL_473] (rows=285117831 width=127) + predicate:((cs_sold_date_sk BETWEEN DynamicValue(RS_90_date_dim_d_date_sk_min) AND DynamicValue(RS_90_date_dim_d_date_sk_max) and in_bloom_filter(cs_sold_date_sk, DynamicValue(RS_90_date_dim_d_date_sk_bloom_filter))) and cs_bill_customer_sk is not null and cs_sold_date_sk is not null) TableScan [TS_0] (rows=287989836 width=127) default@catalog_sales,catalog_sales,Tbl:COMPLETE,Col:COMPLETE,Output:["cs_sold_date_sk","cs_bill_customer_sk","cs_item_sk","cs_quantity","cs_list_price"] <-Reducer 9 [BROADCAST_EDGE] vectorized - BROADCAST [RS_466] - Group By Operator [GBY_465] (rows=1 width=12) + BROADCAST [RS_472] + Group By Operator [GBY_471] (rows=1 width=12) Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] <-Map 8 [CUSTOM_SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_463] - Group By Operator [GBY_461] (rows=1 width=12) + PARTITION_ONLY_SHUFFLE [RS_469] + Group By Operator [GBY_467] (rows=1 width=12) Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] - Select Operator [SEL_458] (rows=50 width=4) + Select Operator [SEL_464] (rows=50 width=4) Output:["_col0"] - Please refer to the previous Select Operator [SEL_456] + Please refer to the previous Select Operator [SEL_462] <-Reducer 20 [SIMPLE_EDGE] - SHUFFLE [RS_92] + SHUFFLE [RS_93] PartitionCols:_col0 Select Operator [SEL_67] (rows=471875 width=3) Output:["_col0"] Filter Operator [FIL_66] (rows=471875 width=227) predicate:(_col3 > (0.95 * _col1)) - Merge Join Operator [MERGEJOIN_438] (rows=1415626 width=227) + Merge Join Operator [MERGEJOIN_444] (rows=1415626 width=227) Conds:(Inner),Output:["_col1","_col2","_col3"] <-Reducer 19 [CUSTOM_SIMPLE_EDGE] PARTITION_ONLY_SHUFFLE [RS_63] - Merge Join Operator [MERGEJOIN_437] (rows=1 width=112) + Merge Join Operator [MERGEJOIN_443] (rows=1 width=112) Conds:(Inner),Output:["_col1"] <-Reducer 18 [CUSTOM_SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_503] - Select Operator [SEL_502] (rows=1 width=8) - Filter Operator [FIL_501] (rows=1 width=8) + PARTITION_ONLY_SHUFFLE [RS_505] + Select Operator [SEL_504] (rows=1 width=8) + Filter Operator [FIL_503] (rows=1 width=8) predicate:(sq_count_check(_col0) <= 1) - Group By Operator [GBY_500] (rows=1 width=8) + Group By Operator [GBY_502] (rows=1 width=8) Output:["_col0"],aggregations:["count()"] - Select Operator [SEL_499] (rows=1 width=8) - Group By Operator [GBY_498] (rows=1 width=8) + Select Operator [SEL_501] (rows=1 width=8) + Group By Operator [GBY_500] (rows=1 width=8) Output:["_col0"],aggregations:["count(VALUE._col0)"] <-Reducer 17 [CUSTOM_SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_494] - Group By Operator [GBY_490] (rows=1 width=8) + PARTITION_ONLY_SHUFFLE [RS_496] + Group By Operator [GBY_492] (rows=1 width=8) Output:["_col0"],aggregations:["count(_col0)"] - Select Operator [SEL_486] (rows=50562 width=112) + Select Operator [SEL_488] (rows=50562 width=112) Output:["_col0"] - Please refer to the previous Group By Operator [GBY_485] + Please refer to the previous Group By Operator [GBY_487] <-Reducer 21 [CUSTOM_SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_505] - Group By Operator [GBY_504] (rows=1 width=112) + PARTITION_ONLY_SHUFFLE [RS_507] + Group By Operator [GBY_506] (rows=1 width=112) Output:["_col0"],aggregations:["max(VALUE._col0)"] <-Reducer 17 [CUSTOM_SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_495] - Group By Operator [GBY_491] (rows=1 width=112) + PARTITION_ONLY_SHUFFLE [RS_497] + Group By Operator [GBY_493] (rows=1 width=112) Output:["_col0"],aggregations:["max(_col1)"] - Select Operator [SEL_487] (rows=50562 width=112) + Select Operator [SEL_489] (rows=50562 width=112) Output:["_col1"] - Please refer to the previous Group By Operator [GBY_485] - <-Reducer 34 [CUSTOM_SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_513] - Group By Operator [GBY_512] (rows=1415626 width=115) + Please refer to the previous Group By Operator [GBY_487] + <-Reducer 29 [CUSTOM_SIMPLE_EDGE] vectorized + PARTITION_ONLY_SHUFFLE [RS_515] + Group By Operator [GBY_514] (rows=1415626 width=115) Output:["_col0","_col1"],aggregations:["sum(VALUE._col0)"],keys:KEY._col0 - <-Map 33 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_511] + <-Map 28 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_513] PartitionCols:_col0 - Group By Operator [GBY_510] (rows=550080312 width=115) + Group By Operator [GBY_512] (rows=550080312 width=115) Output:["_col0","_col1"],aggregations:["sum(_col1)"],keys:_col0 - Select Operator [SEL_509] (rows=550080312 width=114) + Select Operator [SEL_511] (rows=550080312 width=114) Output:["_col0","_col1"] - Filter Operator [FIL_508] (rows=550080312 width=114) - predicate:((ss_customer_sk BETWEEN DynamicValue(RS_91_catalog_sales_cs_bill_customer_sk_min) AND DynamicValue(RS_91_catalog_sales_cs_bill_customer_sk_max) and in_bloom_filter(ss_customer_sk, DynamicValue(RS_91_catalog_sales_cs_bill_customer_sk_bloom_filter))) and ss_customer_sk is not null) + Filter Operator [FIL_510] (rows=550080312 width=114) + predicate:((ss_customer_sk BETWEEN DynamicValue(RS_92_catalog_sales_cs_bill_customer_sk_min) AND DynamicValue(RS_92_catalog_sales_cs_bill_customer_sk_max) and in_bloom_filter(ss_customer_sk, DynamicValue(RS_92_catalog_sales_cs_bill_customer_sk_bloom_filter))) and ss_customer_sk is not null) TableScan [TS_52] (rows=575995635 width=114) default@store_sales,store_sales,Tbl:COMPLETE,Col:COMPLETE,Output:["ss_customer_sk","ss_quantity","ss_sales_price"] <-Reducer 7 [BROADCAST_EDGE] vectorized - BROADCAST [RS_507] - Group By Operator [GBY_506] (rows=1 width=12) + BROADCAST [RS_509] + Group By Operator [GBY_508] (rows=1 width=12) Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] <-Reducer 2 [CUSTOM_SIMPLE_EDGE] - PARTITION_ONLY_SHUFFLE [RS_333] - Group By Operator [GBY_332] (rows=1 width=12) + PARTITION_ONLY_SHUFFLE [RS_339] + Group By Operator [GBY_338] (rows=1 width=12) Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] - Select Operator [SEL_331] (rows=7751875 width=6) + Select Operator [SEL_337] (rows=7751875 width=6) Output:["_col0"] - Please refer to the previous Merge Join Operator [MERGEJOIN_427] - <-Reducer 30 [ONE_TO_ONE_EDGE] vectorized - FORWARD [RS_524] + Please refer to the previous Merge Join Operator [MERGEJOIN_433] + <-Reducer 33 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_534] PartitionCols:_col0 - Select Operator [SEL_523] (rows=745 width=4) - Output:["_col0"] - Filter Operator [FIL_522] (rows=745 width=12) - predicate:(_col1 > 4L) - Group By Operator [GBY_521] (rows=2235 width=12) - Output:["_col0","_col1"],aggregations:["count(VALUE._col0)"],keys:KEY._col0 - <-Reducer 29 [SIMPLE_EDGE] - SHUFFLE [RS_84] - PartitionCols:_col0 - Please refer to the previous Group By Operator [GBY_83] + Group By Operator [GBY_533] (rows=2235 width=4) + Output:["_col0"],keys:_col0 + Select Operator [SEL_532] (rows=6548799 width=4) + Output:["_col0"] + Filter Operator [FIL_531] (rows=6548799 width=290) + predicate:(_col3 > 4L) + Select Operator [SEL_530] (rows=19646398 width=290) + Output:["_col1","_col3"] + Group By Operator [GBY_529] (rows=19646398 width=290) + Output:["_col0","_col1","_col2","_col3"],aggregations:["count(VALUE._col0)"],keys:KEY._col0, KEY._col1, KEY._col2 + <-Reducer 32 [SIMPLE_EDGE] + SHUFFLE [RS_84] + PartitionCols:_col0, _col1, _col2 + Please refer to the previous Group By Operator [GBY_83]
http://git-wip-us.apache.org/repos/asf/hive/blob/83d1fd23/ql/src/test/results/clientpositive/perf/tez/constraints/query39.q.out ---------------------------------------------------------------------- diff --git a/ql/src/test/results/clientpositive/perf/tez/constraints/query39.q.out b/ql/src/test/results/clientpositive/perf/tez/constraints/query39.q.out index 812928b..ca2d830 100644 --- a/ql/src/test/results/clientpositive/perf/tez/constraints/query39.q.out +++ b/ql/src/test/results/clientpositive/perf/tez/constraints/query39.q.out @@ -66,7 +66,7 @@ Vertex dependency in root stage Reducer 2 <- Map 1 (SIMPLE_EDGE), Map 10 (SIMPLE_EDGE) Reducer 3 <- Map 11 (SIMPLE_EDGE), Reducer 2 (SIMPLE_EDGE) Reducer 4 <- Reducer 3 (SIMPLE_EDGE) -Reducer 5 <- Reducer 4 (ONE_TO_ONE_EDGE), Reducer 9 (ONE_TO_ONE_EDGE) +Reducer 5 <- Reducer 4 (SIMPLE_EDGE), Reducer 9 (SIMPLE_EDGE) Reducer 6 <- Reducer 5 (SIMPLE_EDGE) Reducer 7 <- Map 1 (SIMPLE_EDGE), Map 10 (SIMPLE_EDGE) Reducer 8 <- Map 11 (SIMPLE_EDGE), Reducer 7 (SIMPLE_EDGE) @@ -77,99 +77,101 @@ Stage-0 limit:-1 Stage-1 Reducer 6 vectorized - File Output Operator [FS_166] - Select Operator [SEL_165] (rows=859 width=56) + File Output Operator [FS_170] + Select Operator [SEL_169] (rows=189509 width=56) Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6","_col7","_col8","_col9"] <-Reducer 5 [SIMPLE_EDGE] SHUFFLE [RS_44] - Select Operator [SEL_43] (rows=859 width=48) - Output:["_col0","_col1","_col2","_col3","_col6","_col7"] - Merge Join Operator [MERGEJOIN_144] (rows=859 width=40) - Conds:RS_160._col0, _col1=RS_164._col0, _col1(Inner),Output:["_col0","_col1","_col2","_col3","_col6","_col7"] - <-Reducer 4 [ONE_TO_ONE_EDGE] vectorized - FORWARD [RS_160] - PartitionCols:_col0, _col1 - Select Operator [SEL_159] (rows=859 width=24) - Output:["_col0","_col1","_col2","_col3"] - Filter Operator [FIL_158] (rows=859 width=40) - predicate:CASE WHEN (((UDFToDouble(_col2) / _col3) = 0)) THEN (false) ELSE (((power(((_col4 - ((_col5 * _col5) / _col3)) / CASE WHEN ((_col3 = 1L)) THEN (null) ELSE ((_col3 - 1)) END), 0.5) / (UDFToDouble(_col2) / _col3)) > 1.0D)) END - Group By Operator [GBY_157] (rows=1719 width=40) - Output:["_col0","_col1","_col2","_col3","_col4","_col5"],aggregations:["sum(VALUE._col0)","count(VALUE._col1)","sum(VALUE._col2)","sum(VALUE._col3)"],keys:KEY._col0, KEY._col1 + Merge Join Operator [MERGEJOIN_146] (rows=189509 width=48) + Conds:RS_163._col0, _col1=RS_168._col0, _col1(Inner),Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6","_col7"] + <-Reducer 4 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_163] + PartitionCols:_col0, _col1 + Select Operator [SEL_162] (rows=18049 width=24) + Output:["_col0","_col1","_col2","_col3"] + Filter Operator [FIL_161] (rows=18049 width=40) + predicate:CASE WHEN (((UDFToDouble(_col3) / _col4) = 0)) THEN (false) ELSE (((power(((_col5 - ((_col6 * _col6) / _col4)) / CASE WHEN ((_col4 = 1L)) THEN (null) ELSE ((_col4 - 1)) END), 0.5) / (UDFToDouble(_col3) / _col4)) > 1.0D)) END + Select Operator [SEL_160] (rows=36099 width=40) + Output:["_col0","_col1","_col3","_col4","_col5","_col6"] + Group By Operator [GBY_159] (rows=36099 width=140) + Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6"],aggregations:["sum(VALUE._col0)","count(VALUE._col1)","sum(VALUE._col2)","sum(VALUE._col3)"],keys:KEY._col0, KEY._col1, KEY._col2 <-Reducer 3 [SIMPLE_EDGE] SHUFFLE [RS_16] - PartitionCols:_col0, _col1 - Group By Operator [GBY_15] (rows=1719 width=40) - Output:["_col0","_col1","_col2","_col3","_col4","_col5"],aggregations:["sum(_col3)","count(_col3)","sum(_col5)","sum(_col4)"],keys:_col1, _col2 - Select Operator [SEL_13] (rows=1032514 width=8) - Output:["_col1","_col2","_col3","_col4","_col5"] - Merge Join Operator [MERGEJOIN_141] (rows=1032514 width=8) - Conds:RS_10._col1=RS_155._col0(Inner),Output:["_col2","_col3","_col5"] + PartitionCols:_col0, _col1, _col2 + Group By Operator [GBY_15] (rows=36099 width=140) + Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6"],aggregations:["sum(_col3)","count(_col3)","sum(_col5)","sum(_col4)"],keys:_col1, _col2, _col0 + Select Operator [SEL_13] (rows=1032514 width=108) + Output:["_col0","_col1","_col2","_col3","_col4","_col5"] + Merge Join Operator [MERGEJOIN_143] (rows=1032514 width=108) + Conds:RS_10._col1=RS_157._col0(Inner),Output:["_col2","_col3","_col5","_col6"] <-Map 11 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_155] + SHUFFLE [RS_157] PartitionCols:_col0 - Select Operator [SEL_154] (rows=27 width=4) - Output:["_col0"] - TableScan [TS_5] (rows=27 width=4) - default@warehouse,warehouse,Tbl:COMPLETE,Col:COMPLETE,Output:["w_warehouse_sk"] + Select Operator [SEL_156] (rows=27 width=104) + Output:["_col0","_col1"] + TableScan [TS_5] (rows=27 width=104) + default@warehouse,warehouse,Tbl:COMPLETE,Col:COMPLETE,Output:["w_warehouse_sk","w_warehouse_name"] <-Reducer 2 [SIMPLE_EDGE] SHUFFLE [RS_10] PartitionCols:_col1 - Merge Join Operator [MERGEJOIN_140] (rows=1032514 width=8) - Conds:RS_146._col0=RS_152._col0(Inner),Output:["_col1","_col2","_col3"] + Merge Join Operator [MERGEJOIN_142] (rows=1032514 width=8) + Conds:RS_148._col0=RS_154._col0(Inner),Output:["_col1","_col2","_col3"] <-Map 1 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_146] + SHUFFLE [RS_148] PartitionCols:_col0 - Select Operator [SEL_145] (rows=37584000 width=15) + Select Operator [SEL_147] (rows=37584000 width=15) Output:["_col0","_col1","_col2","_col3"] TableScan [TS_0] (rows=37584000 width=15) default@inventory,inventory,Tbl:COMPLETE,Col:COMPLETE,Output:["inv_date_sk","inv_item_sk","inv_warehouse_sk","inv_quantity_on_hand"] <-Map 10 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_152] + SHUFFLE [RS_154] PartitionCols:_col0 - Select Operator [SEL_150] (rows=50 width=4) + Select Operator [SEL_152] (rows=50 width=4) Output:["_col0"] - Filter Operator [FIL_148] (rows=50 width=12) - predicate:((d_moy = 5) and (d_year = 1999)) + Filter Operator [FIL_150] (rows=50 width=12) + predicate:((d_moy = 4) and (d_year = 1999)) TableScan [TS_2] (rows=73049 width=12) default@date_dim,date_dim,Tbl:COMPLETE,Col:COMPLETE,Output:["d_date_sk","d_year","d_moy"] - <-Reducer 9 [ONE_TO_ONE_EDGE] vectorized - FORWARD [RS_164] - PartitionCols:_col0, _col1 - Select Operator [SEL_163] (rows=859 width=24) - Output:["_col0","_col1","_col2","_col3"] - Filter Operator [FIL_162] (rows=859 width=40) - predicate:CASE WHEN (((UDFToDouble(_col2) / _col3) = 0)) THEN (false) ELSE (((power(((_col4 - ((_col5 * _col5) / _col3)) / CASE WHEN ((_col3 = 1L)) THEN (null) ELSE ((_col3 - 1)) END), 0.5) / (UDFToDouble(_col2) / _col3)) > 1.0D)) END - Group By Operator [GBY_161] (rows=1719 width=40) - Output:["_col0","_col1","_col2","_col3","_col4","_col5"],aggregations:["sum(VALUE._col0)","count(VALUE._col1)","sum(VALUE._col2)","sum(VALUE._col3)"],keys:KEY._col0, KEY._col1 + <-Reducer 9 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_168] + PartitionCols:_col0, _col1 + Select Operator [SEL_167] (rows=18049 width=24) + Output:["_col0","_col1","_col2","_col3"] + Filter Operator [FIL_166] (rows=18049 width=40) + predicate:CASE WHEN (((UDFToDouble(_col3) / _col4) = 0)) THEN (false) ELSE (((power(((_col5 - ((_col6 * _col6) / _col4)) / CASE WHEN ((_col4 = 1L)) THEN (null) ELSE ((_col4 - 1)) END), 0.5) / (UDFToDouble(_col3) / _col4)) > 1.0D)) END + Select Operator [SEL_165] (rows=36099 width=40) + Output:["_col0","_col1","_col3","_col4","_col5","_col6"] + Group By Operator [GBY_164] (rows=36099 width=140) + Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6"],aggregations:["sum(VALUE._col0)","count(VALUE._col1)","sum(VALUE._col2)","sum(VALUE._col3)"],keys:KEY._col0, KEY._col1, KEY._col2 <-Reducer 8 [SIMPLE_EDGE] SHUFFLE [RS_36] - PartitionCols:_col0, _col1 - Group By Operator [GBY_35] (rows=1719 width=40) - Output:["_col0","_col1","_col2","_col3","_col4","_col5"],aggregations:["sum(_col3)","count(_col3)","sum(_col5)","sum(_col4)"],keys:_col1, _col2 - Select Operator [SEL_33] (rows=1032514 width=8) - Output:["_col1","_col2","_col3","_col4","_col5"] - Merge Join Operator [MERGEJOIN_143] (rows=1032514 width=8) - Conds:RS_30._col1=RS_156._col0(Inner),Output:["_col2","_col3","_col5"] + PartitionCols:_col0, _col1, _col2 + Group By Operator [GBY_35] (rows=36099 width=140) + Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6"],aggregations:["sum(_col3)","count(_col3)","sum(_col5)","sum(_col4)"],keys:_col1, _col2, _col0 + Select Operator [SEL_33] (rows=1032514 width=108) + Output:["_col0","_col1","_col2","_col3","_col4","_col5"] + Merge Join Operator [MERGEJOIN_145] (rows=1032514 width=108) + Conds:RS_30._col1=RS_158._col0(Inner),Output:["_col2","_col3","_col5","_col6"] <-Map 11 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_156] + SHUFFLE [RS_158] PartitionCols:_col0 - Please refer to the previous Select Operator [SEL_154] + Please refer to the previous Select Operator [SEL_156] <-Reducer 7 [SIMPLE_EDGE] SHUFFLE [RS_30] PartitionCols:_col1 - Merge Join Operator [MERGEJOIN_142] (rows=1032514 width=8) - Conds:RS_147._col0=RS_153._col0(Inner),Output:["_col1","_col2","_col3"] + Merge Join Operator [MERGEJOIN_144] (rows=1032514 width=8) + Conds:RS_149._col0=RS_155._col0(Inner),Output:["_col1","_col2","_col3"] <-Map 1 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_147] + SHUFFLE [RS_149] PartitionCols:_col0 - Please refer to the previous Select Operator [SEL_145] + Please refer to the previous Select Operator [SEL_147] <-Map 10 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_153] + SHUFFLE [RS_155] PartitionCols:_col0 - Select Operator [SEL_151] (rows=50 width=4) + Select Operator [SEL_153] (rows=50 width=4) Output:["_col0"] - Filter Operator [FIL_149] (rows=50 width=12) - predicate:((d_moy = 4) and (d_year = 1999)) + Filter Operator [FIL_151] (rows=50 width=12) + predicate:((d_moy = 5) and (d_year = 1999)) Please refer to the previous TableScan [TS_2] PREHOOK: query: with inv as http://git-wip-us.apache.org/repos/asf/hive/blob/83d1fd23/ql/src/test/results/clientpositive/perf/tez/constraints/query94.q.out ---------------------------------------------------------------------- diff --git a/ql/src/test/results/clientpositive/perf/tez/constraints/query94.q.out b/ql/src/test/results/clientpositive/perf/tez/constraints/query94.q.out index 7290441..1dba4fb 100644 --- a/ql/src/test/results/clientpositive/perf/tez/constraints/query94.q.out +++ b/ql/src/test/results/clientpositive/perf/tez/constraints/query94.q.out @@ -73,11 +73,12 @@ Map 1 <- Reducer 13 (BROADCAST_EDGE) Map 15 <- Reducer 10 (BROADCAST_EDGE) Reducer 10 <- Reducer 4 (CUSTOM_SIMPLE_EDGE) Reducer 13 <- Map 12 (CUSTOM_SIMPLE_EDGE) +Reducer 17 <- Map 16 (SIMPLE_EDGE) Reducer 2 <- Map 1 (SIMPLE_EDGE), Map 11 (SIMPLE_EDGE) Reducer 3 <- Map 12 (SIMPLE_EDGE), Reducer 2 (SIMPLE_EDGE) Reducer 4 <- Map 14 (SIMPLE_EDGE), Reducer 3 (SIMPLE_EDGE) Reducer 5 <- Map 15 (SIMPLE_EDGE), Reducer 4 (SIMPLE_EDGE) -Reducer 6 <- Map 16 (SIMPLE_EDGE), Reducer 5 (ONE_TO_ONE_EDGE) +Reducer 6 <- Reducer 17 (ONE_TO_ONE_EDGE), Reducer 5 (ONE_TO_ONE_EDGE) Reducer 7 <- Reducer 6 (SIMPLE_EDGE) Reducer 8 <- Reducer 7 (CUSTOM_SIMPLE_EDGE) Reducer 9 <- Reducer 8 (SIMPLE_EDGE) @@ -87,130 +88,137 @@ Stage-0 limit:-1 Stage-1 Reducer 9 vectorized - File Output Operator [FS_159] - Limit [LIM_158] (rows=1 width=240) + File Output Operator [FS_166] + Limit [LIM_165] (rows=1 width=240) Number of rows:100 - Select Operator [SEL_157] (rows=1 width=240) + Select Operator [SEL_164] (rows=1 width=240) Output:["_col0","_col1","_col2"] <-Reducer 8 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_156] - Select Operator [SEL_155] (rows=1 width=240) + SHUFFLE [RS_163] + Select Operator [SEL_162] (rows=1 width=240) Output:["_col1","_col2","_col3"] - Group By Operator [GBY_154] (rows=1 width=232) + Group By Operator [GBY_161] (rows=1 width=232) Output:["_col0","_col1","_col2"],aggregations:["count(VALUE._col0)","sum(VALUE._col1)","sum(VALUE._col2)"] <-Reducer 7 [CUSTOM_SIMPLE_EDGE] vectorized - PARTITION_ONLY_SHUFFLE [RS_153] - Group By Operator [GBY_152] (rows=1 width=232) + PARTITION_ONLY_SHUFFLE [RS_160] + Group By Operator [GBY_159] (rows=1 width=232) Output:["_col0","_col1","_col2"],aggregations:["count(_col0)","sum(_col1)","sum(_col2)"] - Group By Operator [GBY_151] (rows=2511437 width=228) + Group By Operator [GBY_158] (rows=2511437 width=228) Output:["_col0","_col1","_col2"],aggregations:["sum(VALUE._col0)","sum(VALUE._col1)"],keys:KEY._col0 <-Reducer 6 [SIMPLE_EDGE] - SHUFFLE [RS_69] + SHUFFLE [RS_73] PartitionCols:_col0 - Group By Operator [GBY_68] (rows=2511437 width=228) + Group By Operator [GBY_72] (rows=2511437 width=228) Output:["_col0","_col2","_col3"],aggregations:["sum(_col5)","sum(_col6)"],keys:_col4 - Select Operator [SEL_37] (rows=5022875 width=230) + Select Operator [SEL_41] (rows=5022875 width=229) Output:["_col4","_col5","_col6"] - Filter Operator [FIL_36] (rows=5022875 width=230) + Filter Operator [FIL_40] (rows=5022875 width=229) predicate:_col14 is null - Merge Join Operator [MERGEJOIN_125] (rows=14054072 width=230) - Conds:RS_33._col4=RS_150._col0(Left Outer),Output:["_col4","_col5","_col6","_col14"] - <-Map 16 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_150] + Merge Join Operator [MERGEJOIN_129] (rows=10045750 width=229) + Conds:RS_37._col4=RS_157._col0(Left Outer),Output:["_col4","_col5","_col6","_col14"] + <-Reducer 17 [ONE_TO_ONE_EDGE] vectorized + FORWARD [RS_157] PartitionCols:_col0 - Select Operator [SEL_149] (rows=14398467 width=8) + Select Operator [SEL_156] (rows=8007986 width=8) Output:["_col0","_col1"] - TableScan [TS_25] (rows=14398467 width=4) - default@web_returns,wr1,Tbl:COMPLETE,Col:COMPLETE,Output:["wr_order_number"] + Group By Operator [GBY_155] (rows=8007986 width=4) + Output:["_col0"],keys:KEY._col0 + <-Map 16 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_154] + PartitionCols:_col0 + Group By Operator [GBY_153] (rows=14398467 width=4) + Output:["_col0"],keys:wr_order_number + TableScan [TS_25] (rows=14398467 width=4) + default@web_returns,wr1,Tbl:COMPLETE,Col:COMPLETE,Output:["wr_order_number"] <-Reducer 5 [ONE_TO_ONE_EDGE] - FORWARD [RS_33] + FORWARD [RS_37] PartitionCols:_col4 - Select Operator [SEL_32] (rows=5022875 width=231) + Select Operator [SEL_36] (rows=5022875 width=231) Output:["_col4","_col5","_col6"] - Merge Join Operator [MERGEJOIN_124] (rows=5022875 width=235) - Conds:RS_29._col4=RS_148._col0(Left Semi),Output:["_col3","_col4","_col5","_col6","_col14"],residual filter predicates:{(_col3 <> _col14)} + Merge Join Operator [MERGEJOIN_128] (rows=5022875 width=235) + Conds:RS_33._col4=RS_152._col0(Left Semi),Output:["_col3","_col4","_col5","_col6","_col14"],residual filter predicates:{(_col3 <> _col14)} <-Reducer 4 [SIMPLE_EDGE] - SHUFFLE [RS_29] + SHUFFLE [RS_33] PartitionCols:_col4 - Merge Join Operator [MERGEJOIN_123] (rows=5022875 width=231) - Conds:RS_18._col2=RS_142._col0(Inner),Output:["_col3","_col4","_col5","_col6"] + Merge Join Operator [MERGEJOIN_127] (rows=5022875 width=231) + Conds:RS_18._col2=RS_146._col0(Inner),Output:["_col3","_col4","_col5","_col6"] <-Map 14 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_142] + SHUFFLE [RS_146] PartitionCols:_col0 - Select Operator [SEL_141] (rows=12 width=91) + Select Operator [SEL_145] (rows=12 width=91) Output:["_col0"] - Filter Operator [FIL_140] (rows=12 width=92) + Filter Operator [FIL_144] (rows=12 width=92) predicate:(web_company_name = 'pri') TableScan [TS_9] (rows=84 width=92) default@web_site,web_site,Tbl:COMPLETE,Col:COMPLETE,Output:["web_site_sk","web_company_name"] <-Reducer 3 [SIMPLE_EDGE] SHUFFLE [RS_18] PartitionCols:_col2 - Merge Join Operator [MERGEJOIN_122] (rows=15673790 width=235) - Conds:RS_15._col1=RS_128._col0(Inner),Output:["_col2","_col3","_col4","_col5","_col6"] + Merge Join Operator [MERGEJOIN_126] (rows=15673790 width=235) + Conds:RS_15._col1=RS_132._col0(Inner),Output:["_col2","_col3","_col4","_col5","_col6"] <-Map 12 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_128] + SHUFFLE [RS_132] PartitionCols:_col0 - Select Operator [SEL_127] (rows=784314 width=90) + Select Operator [SEL_131] (rows=784314 width=90) Output:["_col0"] - Filter Operator [FIL_126] (rows=784314 width=90) + Filter Operator [FIL_130] (rows=784314 width=90) predicate:(ca_state = 'TX') TableScan [TS_6] (rows=40000000 width=90) default@customer_address,customer_address,Tbl:COMPLETE,Col:COMPLETE,Output:["ca_address_sk","ca_state"] <-Reducer 2 [SIMPLE_EDGE] SHUFFLE [RS_15] PartitionCols:_col1 - Merge Join Operator [MERGEJOIN_121] (rows=15987241 width=239) - Conds:RS_136._col0=RS_139._col0(Inner),Output:["_col1","_col2","_col3","_col4","_col5","_col6"] + Merge Join Operator [MERGEJOIN_125] (rows=15987241 width=239) + Conds:RS_140._col0=RS_143._col0(Inner),Output:["_col1","_col2","_col3","_col4","_col5","_col6"] <-Map 1 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_136] + SHUFFLE [RS_140] PartitionCols:_col0 - Select Operator [SEL_135] (rows=143895019 width=243) + Select Operator [SEL_139] (rows=143895019 width=243) Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6"] - Filter Operator [FIL_134] (rows=143895019 width=243) + Filter Operator [FIL_138] (rows=143895019 width=243) predicate:((ws_ship_addr_sk BETWEEN DynamicValue(RS_16_customer_address_ca_address_sk_min) AND DynamicValue(RS_16_customer_address_ca_address_sk_max) and in_bloom_filter(ws_ship_addr_sk, DynamicValue(RS_16_customer_address_ca_address_sk_bloom_filter))) and ws_ship_addr_sk is not null and ws_ship_date_sk is not null and ws_web_site_sk is not null) TableScan [TS_0] (rows=144002668 width=243) default@web_sales,ws1,Tbl:COMPLETE,Col:COMPLETE,Output:["ws_ship_date_sk","ws_ship_addr_sk","ws_web_site_sk","ws_warehouse_sk","ws_order_number","ws_ext_ship_cost","ws_net_profit"] <-Reducer 13 [BROADCAST_EDGE] vectorized - BROADCAST [RS_133] - Group By Operator [GBY_132] (rows=1 width=12) + BROADCAST [RS_137] + Group By Operator [GBY_136] (rows=1 width=12) Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] <-Map 12 [CUSTOM_SIMPLE_EDGE] vectorized - SHUFFLE [RS_131] - Group By Operator [GBY_130] (rows=1 width=12) + SHUFFLE [RS_135] + Group By Operator [GBY_134] (rows=1 width=12) Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] - Select Operator [SEL_129] (rows=784314 width=4) + Select Operator [SEL_133] (rows=784314 width=4) Output:["_col0"] - Please refer to the previous Select Operator [SEL_127] + Please refer to the previous Select Operator [SEL_131] <-Map 11 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_139] + SHUFFLE [RS_143] PartitionCols:_col0 - Select Operator [SEL_138] (rows=8116 width=98) + Select Operator [SEL_142] (rows=8116 width=98) Output:["_col0"] - Filter Operator [FIL_137] (rows=8116 width=98) + Filter Operator [FIL_141] (rows=8116 width=98) predicate:CAST( d_date AS TIMESTAMP) BETWEEN TIMESTAMP'1999-05-01 00:00:00' AND TIMESTAMP'1999-06-30 00:00:00' TableScan [TS_3] (rows=73049 width=98) default@date_dim,date_dim,Tbl:COMPLETE,Col:COMPLETE,Output:["d_date_sk","d_date"] <-Map 15 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_148] + SHUFFLE [RS_152] PartitionCols:_col0 - Group By Operator [GBY_147] (rows=143966743 width=7) + Group By Operator [GBY_151] (rows=143966743 width=7) Output:["_col0","_col1"],keys:_col0, _col1 - Select Operator [SEL_146] (rows=143966743 width=7) + Select Operator [SEL_150] (rows=143966743 width=7) Output:["_col0","_col1"] - Filter Operator [FIL_145] (rows=143966743 width=7) - predicate:((ws_order_number BETWEEN DynamicValue(RS_29_ws1_ws_order_number_min) AND DynamicValue(RS_29_ws1_ws_order_number_max) and in_bloom_filter(ws_order_number, DynamicValue(RS_29_ws1_ws_order_number_bloom_filter))) and ws_warehouse_sk is not null) + Filter Operator [FIL_149] (rows=143966743 width=7) + predicate:((ws_order_number BETWEEN DynamicValue(RS_33_ws1_ws_order_number_min) AND DynamicValue(RS_33_ws1_ws_order_number_max) and in_bloom_filter(ws_order_number, DynamicValue(RS_33_ws1_ws_order_number_bloom_filter))) and ws_warehouse_sk is not null) TableScan [TS_22] (rows=144002668 width=7) default@web_sales,ws2,Tbl:COMPLETE,Col:COMPLETE,Output:["ws_warehouse_sk","ws_order_number"] <-Reducer 10 [BROADCAST_EDGE] vectorized - BROADCAST [RS_144] - Group By Operator [GBY_143] (rows=1 width=12) + BROADCAST [RS_148] + Group By Operator [GBY_147] (rows=1 width=12) Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] <-Reducer 4 [CUSTOM_SIMPLE_EDGE] - SHUFFLE [RS_111] - Group By Operator [GBY_110] (rows=1 width=12) + SHUFFLE [RS_115] + Group By Operator [GBY_114] (rows=1 width=12) Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] - Select Operator [SEL_109] (rows=5022875 width=8) + Select Operator [SEL_113] (rows=5022875 width=8) Output:["_col0"] - Please refer to the previous Merge Join Operator [MERGEJOIN_123] + Please refer to the previous Merge Join Operator [MERGEJOIN_127]
