http://git-wip-us.apache.org/repos/asf/hive/blob/f2c4f319/ql/src/test/results/clientpositive/perf/tez/query5.q.out ---------------------------------------------------------------------- diff --git a/ql/src/test/results/clientpositive/perf/tez/query5.q.out b/ql/src/test/results/clientpositive/perf/tez/query5.q.out index f074ed1..340f8b1 100644 --- a/ql/src/test/results/clientpositive/perf/tez/query5.q.out +++ b/ql/src/test/results/clientpositive/perf/tez/query5.q.out @@ -255,20 +255,29 @@ POSTHOOK: type: QUERY Plan optimized by CBO. Vertex dependency in root stage -Map 1 <- Union 2 (CONTAINS) -Map 18 <- Union 19 (CONTAINS) -Map 20 <- Union 19 (CONTAINS) -Map 22 <- Union 23 (CONTAINS) +Map 1 <- Reducer 11 (BROADCAST_EDGE), Reducer 21 (BROADCAST_EDGE), Union 2 (CONTAINS) +Map 22 <- Reducer 15 (BROADCAST_EDGE), Reducer 26 (BROADCAST_EDGE), Union 23 (CONTAINS) +Map 24 <- Union 23 (CONTAINS) +Map 27 <- Reducer 19 (BROADCAST_EDGE), Reducer 35 (BROADCAST_EDGE), Union 28 (CONTAINS) +Map 29 <- Reducer 32 (BROADCAST_EDGE), Reducer 33 (BROADCAST_EDGE), Reducer 35 (BROADCAST_EDGE) Map 9 <- Union 2 (CONTAINS) -Reducer 11 <- Map 10 (SIMPLE_EDGE), Union 19 (SIMPLE_EDGE) -Reducer 12 <- Map 21 (SIMPLE_EDGE), Reducer 11 (SIMPLE_EDGE) -Reducer 13 <- Reducer 12 (SIMPLE_EDGE), Union 6 (CONTAINS) -Reducer 14 <- Map 10 (SIMPLE_EDGE), Union 23 (SIMPLE_EDGE) -Reducer 15 <- Map 27 (SIMPLE_EDGE), Reducer 14 (SIMPLE_EDGE) -Reducer 16 <- Reducer 15 (SIMPLE_EDGE), Union 6 (CONTAINS) -Reducer 25 <- Map 24 (SIMPLE_EDGE), Map 26 (SIMPLE_EDGE), Union 23 (CONTAINS) +Reducer 11 <- Map 10 (CUSTOM_SIMPLE_EDGE) +Reducer 12 <- Map 10 (SIMPLE_EDGE), Union 23 (SIMPLE_EDGE) +Reducer 13 <- Map 25 (SIMPLE_EDGE), Reducer 12 (SIMPLE_EDGE) +Reducer 14 <- Reducer 13 (SIMPLE_EDGE), Union 6 (CONTAINS) +Reducer 15 <- Map 10 (CUSTOM_SIMPLE_EDGE) +Reducer 16 <- Map 10 (SIMPLE_EDGE), Union 28 (SIMPLE_EDGE) +Reducer 17 <- Map 34 (SIMPLE_EDGE), Reducer 16 (SIMPLE_EDGE) +Reducer 18 <- Reducer 17 (SIMPLE_EDGE), Union 6 (CONTAINS) +Reducer 19 <- Map 10 (CUSTOM_SIMPLE_EDGE) +Reducer 21 <- Map 20 (CUSTOM_SIMPLE_EDGE) +Reducer 26 <- Map 25 (CUSTOM_SIMPLE_EDGE) Reducer 3 <- Map 10 (SIMPLE_EDGE), Union 2 (SIMPLE_EDGE) -Reducer 4 <- Map 17 (SIMPLE_EDGE), Reducer 3 (SIMPLE_EDGE) +Reducer 30 <- Map 29 (SIMPLE_EDGE), Map 31 (SIMPLE_EDGE), Union 28 (CONTAINS) +Reducer 32 <- Map 31 (CUSTOM_SIMPLE_EDGE) +Reducer 33 <- Map 31 (CUSTOM_SIMPLE_EDGE) +Reducer 35 <- Map 34 (CUSTOM_SIMPLE_EDGE) +Reducer 4 <- Map 20 (SIMPLE_EDGE), Reducer 3 (SIMPLE_EDGE) Reducer 5 <- Reducer 4 (SIMPLE_EDGE), Union 6 (CONTAINS) Reducer 7 <- Union 6 (SIMPLE_EDGE) Reducer 8 <- Reducer 7 (SIMPLE_EDGE) @@ -278,196 +287,287 @@ Stage-0 limit:100 Stage-1 Reducer 8 vectorized - File Output Operator [FS_202] - Limit [LIM_201] (rows=100 width=110) + File Output Operator [FS_304] + Limit [LIM_303] (rows=100 width=110) Number of rows:100 - Select Operator [SEL_200] (rows=1136898901 width=110) + Select Operator [SEL_302] (rows=1136898901 width=110) Output:["_col0","_col1","_col2","_col3","_col4"] <-Reducer 7 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_199] - Select Operator [SEL_198] (rows=1136898901 width=110) + SHUFFLE [RS_301] + Select Operator [SEL_300] (rows=1136898901 width=110) Output:["_col0","_col1","_col2","_col3","_col4"] - Group By Operator [GBY_197] (rows=1136898901 width=110) + Group By Operator [GBY_299] (rows=1136898901 width=110) Output:["_col0","_col1","_col3","_col4","_col5"],aggregations:["sum(VALUE._col0)","sum(VALUE._col1)","sum(VALUE._col2)"],keys:KEY._col0, KEY._col1, KEY._col2 <-Union 6 [SIMPLE_EDGE] - <-Reducer 13 [CONTAINS] vectorized - Reduce Output Operator [RS_212] + <-Reducer 14 [CONTAINS] vectorized + Reduce Output Operator [RS_317] PartitionCols:_col0, _col1, _col2 - Group By Operator [GBY_211] (rows=2273797803 width=110) + Group By Operator [GBY_316] (rows=2273797803 width=110) Output:["_col0","_col1","_col2","_col3","_col4","_col5"],aggregations:["sum(_col2)","sum(_col3)","sum(_col4)"],keys:_col0, _col1, 0L - Select Operator [SEL_210] (rows=191657181 width=132) + Select Operator [SEL_315] (rows=191657181 width=132) Output:["_col0","_col1","_col2","_col3","_col4"] - Group By Operator [GBY_209] (rows=191657181 width=132) + Group By Operator [GBY_314] (rows=191657181 width=132) Output:["_col0","_col1","_col2","_col3","_col4"],aggregations:["sum(VALUE._col0)","sum(VALUE._col1)","sum(VALUE._col2)","sum(VALUE._col3)"],keys:KEY._col0 - <-Reducer 12 [SIMPLE_EDGE] + <-Reducer 13 [SIMPLE_EDGE] SHUFFLE [RS_47] PartitionCols:_col0 Group By Operator [GBY_46] (rows=383314363 width=132) Output:["_col0","_col1","_col2","_col3","_col4"],aggregations:["sum(_col2)","sum(_col4)","sum(_col3)","sum(_col5)"],keys:_col9 - Merge Join Operator [MERGEJOIN_135] (rows=383314363 width=132) - Conds:RS_42._col0=RS_208._col0(Inner),Output:["_col2","_col3","_col4","_col5","_col9"] - <-Map 21 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_208] + Merge Join Operator [MERGEJOIN_221] (rows=383314363 width=132) + Conds:RS_42._col0=RS_310._col0(Inner),Output:["_col2","_col3","_col4","_col5","_col9"] + <-Map 25 [SIMPLE_EDGE] vectorized + PARTITION_ONLY_SHUFFLE [RS_310] PartitionCols:_col0 - Select Operator [SEL_207] (rows=46000 width=460) + Select Operator [SEL_309] (rows=46000 width=460) Output:["_col0","_col1"] - Filter Operator [FIL_206] (rows=46000 width=460) + Filter Operator [FIL_308] (rows=46000 width=460) predicate:cp_catalog_page_sk is not null TableScan [TS_36] (rows=46000 width=460) default@catalog_page,catalog_page,Tbl:COMPLETE,Col:NONE,Output:["cp_catalog_page_sk","cp_catalog_page_id"] - <-Reducer 11 [SIMPLE_EDGE] + <-Reducer 12 [SIMPLE_EDGE] SHUFFLE [RS_42] PartitionCols:_col0 - Merge Join Operator [MERGEJOIN_134] (rows=348467596 width=132) - Conds:Union 19._col1=RS_188._col0(Inner),Output:["_col0","_col2","_col3","_col4","_col5"] + Merge Join Operator [MERGEJOIN_220] (rows=348467596 width=132) + Conds:Union 23._col1=RS_272._col0(Inner),Output:["_col0","_col2","_col3","_col4","_col5"] <-Map 10 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_188] + SHUFFLE [RS_272] PartitionCols:_col0 - Select Operator [SEL_186] (rows=8116 width=1119) + Select Operator [SEL_269] (rows=8116 width=1119) Output:["_col0"] - Filter Operator [FIL_185] (rows=8116 width=1119) + Filter Operator [FIL_268] (rows=8116 width=1119) predicate:(CAST( d_date AS TIMESTAMP) BETWEEN TIMESTAMP'1998-08-04 00:00:00.0' AND TIMESTAMP'1998-08-18 00:00:00.0' and d_date_sk is not null) TableScan [TS_8] (rows=73049 width=1119) default@date_dim,date_dim,Tbl:COMPLETE,Col:NONE,Output:["d_date_sk","d_date"] - <-Union 19 [SIMPLE_EDGE] - <-Map 18 [CONTAINS] vectorized - Reduce Output Operator [RS_222] + <-Union 23 [SIMPLE_EDGE] + <-Map 22 [CONTAINS] vectorized + Reduce Output Operator [RS_334] PartitionCols:_col1 - Select Operator [SEL_221] (rows=287989836 width=135) + Select Operator [SEL_333] (rows=287989836 width=135) Output:["_col0","_col1","_col2","_col3","_col4","_col5"] - Filter Operator [FIL_220] (rows=287989836 width=135) - predicate:(cs_catalog_page_sk is not null and cs_sold_date_sk is not null) - TableScan [TS_163] (rows=287989836 width=135) + Filter Operator [FIL_332] (rows=287989836 width=135) + predicate:((cs_catalog_page_sk BETWEEN DynamicValue(RS_43_catalog_page_cp_catalog_page_sk_min) AND DynamicValue(RS_43_catalog_page_cp_catalog_page_sk_max) and in_bloom_filter(cs_catalog_page_sk, DynamicValue(RS_43_catalog_page_cp_catalog_page_sk_bloom_filter))) and (cs_sold_date_sk BETWEEN DynamicValue(RS_40_date_dim_d_date_sk_min) AND DynamicValue(RS_40_date_dim_d_date_sk_max) and in_bloom_filter(cs_sold_date_sk, DynamicValue(RS_40_date_dim_d_date_sk_bloom_filter))) and cs_catalog_page_sk is not null and cs_sold_date_sk is not null) + TableScan [TS_249] (rows=287989836 width=135) Output:["cs_sold_date_sk","cs_catalog_page_sk","cs_ext_sales_price","cs_net_profit"] - <-Map 20 [CONTAINS] vectorized - Reduce Output Operator [RS_225] + <-Reducer 15 [BROADCAST_EDGE] vectorized + BROADCAST [RS_329] + Group By Operator [GBY_328] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] + <-Map 10 [CUSTOM_SIMPLE_EDGE] vectorized + SHUFFLE [RS_280] + Group By Operator [GBY_277] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] + Select Operator [SEL_273] (rows=8116 width=1119) + Output:["_col0"] + Please refer to the previous Select Operator [SEL_269] + <-Reducer 26 [BROADCAST_EDGE] vectorized + BROADCAST [RS_331] + Group By Operator [GBY_330] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] + <-Map 25 [CUSTOM_SIMPLE_EDGE] vectorized + PARTITION_ONLY_SHUFFLE [RS_313] + Group By Operator [GBY_312] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] + Select Operator [SEL_311] (rows=46000 width=460) + Output:["_col0"] + Please refer to the previous Select Operator [SEL_309] + <-Map 24 [CONTAINS] vectorized + Reduce Output Operator [RS_337] PartitionCols:_col1 - Select Operator [SEL_224] (rows=28798881 width=106) + Select Operator [SEL_336] (rows=28798881 width=106) Output:["_col0","_col1","_col2","_col3","_col4","_col5"] - Filter Operator [FIL_223] (rows=28798881 width=106) + Filter Operator [FIL_335] (rows=28798881 width=106) predicate:(cr_catalog_page_sk is not null and cr_returned_date_sk is not null) - TableScan [TS_168] (rows=28798881 width=106) + TableScan [TS_254] (rows=28798881 width=106) Output:["cr_returned_date_sk","cr_catalog_page_sk","cr_return_amount","cr_net_loss"] - <-Reducer 16 [CONTAINS] vectorized - Reduce Output Operator [RS_219] + <-Reducer 18 [CONTAINS] vectorized + Reduce Output Operator [RS_327] PartitionCols:_col0, _col1, _col2 - Group By Operator [GBY_218] (rows=2273797803 width=110) + Group By Operator [GBY_326] (rows=2273797803 width=110) Output:["_col0","_col1","_col2","_col3","_col4","_col5"],aggregations:["sum(_col2)","sum(_col3)","sum(_col4)"],keys:_col0, _col1, 0L - Select Operator [SEL_217] (rows=182955399 width=135) + Select Operator [SEL_325] (rows=182955399 width=135) Output:["_col0","_col1","_col2","_col3","_col4"] - Group By Operator [GBY_216] (rows=182955399 width=135) + Group By Operator [GBY_324] (rows=182955399 width=135) Output:["_col0","_col1","_col2","_col3","_col4"],aggregations:["sum(VALUE._col0)","sum(VALUE._col1)","sum(VALUE._col2)","sum(VALUE._col3)"],keys:KEY._col0 - <-Reducer 15 [SIMPLE_EDGE] + <-Reducer 17 [SIMPLE_EDGE] SHUFFLE [RS_80] PartitionCols:_col0 Group By Operator [GBY_79] (rows=365910798 width=135) Output:["_col0","_col1","_col2","_col3","_col4"],aggregations:["sum(_col2)","sum(_col4)","sum(_col3)","sum(_col5)"],keys:_col9 - Merge Join Operator [MERGEJOIN_137] (rows=365910798 width=135) - Conds:RS_75._col0=RS_215._col0(Inner),Output:["_col2","_col3","_col4","_col5","_col9"] - <-Map 27 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_215] + Merge Join Operator [MERGEJOIN_223] (rows=365910798 width=135) + Conds:RS_75._col0=RS_320._col0(Inner),Output:["_col2","_col3","_col4","_col5","_col9"] + <-Map 34 [SIMPLE_EDGE] vectorized + PARTITION_ONLY_SHUFFLE [RS_320] PartitionCols:_col0 - Select Operator [SEL_214] (rows=84 width=1850) + Select Operator [SEL_319] (rows=84 width=1850) Output:["_col0","_col1"] - Filter Operator [FIL_213] (rows=84 width=1850) + Filter Operator [FIL_318] (rows=84 width=1850) predicate:web_site_sk is not null TableScan [TS_69] (rows=84 width=1850) default@web_site,web_site,Tbl:COMPLETE,Col:NONE,Output:["web_site_sk","web_site_id"] - <-Reducer 14 [SIMPLE_EDGE] + <-Reducer 16 [SIMPLE_EDGE] SHUFFLE [RS_75] PartitionCols:_col0 - Merge Join Operator [MERGEJOIN_136] (rows=332646173 width=135) - Conds:Union 23._col1=RS_189._col0(Inner),Output:["_col0","_col2","_col3","_col4","_col5"] + Merge Join Operator [MERGEJOIN_222] (rows=332646173 width=135) + Conds:Union 28._col1=RS_274._col0(Inner),Output:["_col0","_col2","_col3","_col4","_col5"] <-Map 10 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_189] + SHUFFLE [RS_274] PartitionCols:_col0 - Please refer to the previous Select Operator [SEL_186] - <-Union 23 [SIMPLE_EDGE] - <-Map 22 [CONTAINS] vectorized - Reduce Output Operator [RS_228] + Please refer to the previous Select Operator [SEL_269] + <-Union 28 [SIMPLE_EDGE] + <-Map 27 [CONTAINS] vectorized + Reduce Output Operator [RS_345] PartitionCols:_col1 - Select Operator [SEL_227] (rows=144002668 width=135) + Select Operator [SEL_344] (rows=144002668 width=135) Output:["_col0","_col1","_col2","_col3","_col4","_col5"] - Filter Operator [FIL_226] (rows=144002668 width=135) - predicate:(ws_sold_date_sk is not null and ws_web_site_sk is not null) - TableScan [TS_173] (rows=144002668 width=135) + Filter Operator [FIL_343] (rows=144002668 width=135) + predicate:((ws_sold_date_sk BETWEEN DynamicValue(RS_73_date_dim_d_date_sk_min) AND DynamicValue(RS_73_date_dim_d_date_sk_max) and in_bloom_filter(ws_sold_date_sk, DynamicValue(RS_73_date_dim_d_date_sk_bloom_filter))) and (ws_web_site_sk BETWEEN DynamicValue(RS_76_web_site_web_site_sk_min) AND DynamicValue(RS_76_web_site_web_site_sk_max) and in_bloom_filter(ws_web_site_sk, DynamicValue(RS_76_web_site_web_site_sk_bloom_filter))) and ws_sold_date_sk is not null and ws_web_site_sk is not null) + TableScan [TS_259] (rows=144002668 width=135) Output:["ws_sold_date_sk","ws_web_site_sk","ws_ext_sales_price","ws_net_profit"] - <-Reducer 25 [CONTAINS] - Reduce Output Operator [RS_181] + <-Reducer 35 [BROADCAST_EDGE] vectorized + BROADCAST [RS_341] + Group By Operator [GBY_340] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] + <-Map 34 [CUSTOM_SIMPLE_EDGE] vectorized + PARTITION_ONLY_SHUFFLE [RS_323] + Group By Operator [GBY_322] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] + Select Operator [SEL_321] (rows=84 width=1850) + Output:["_col0"] + Please refer to the previous Select Operator [SEL_319] + <-Reducer 19 [BROADCAST_EDGE] vectorized + BROADCAST [RS_339] + Group By Operator [GBY_338] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] + <-Map 10 [CUSTOM_SIMPLE_EDGE] vectorized + SHUFFLE [RS_281] + Group By Operator [GBY_278] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] + Select Operator [SEL_275] (rows=8116 width=1119) + Output:["_col0"] + Please refer to the previous Select Operator [SEL_269] + <-Reducer 30 [CONTAINS] + Reduce Output Operator [RS_267] PartitionCols:_col1 - Select Operator [SEL_179] (rows=158402938 width=135) + Select Operator [SEL_265] (rows=158402938 width=135) Output:["_col0","_col1","_col2","_col3","_col4","_col5"] - Merge Join Operator [MERGEJOIN_178] (rows=158402938 width=135) - Conds:RS_231._col0, _col2=RS_234._col1, _col2(Inner),Output:["_col1","_col3","_col6","_col7"] - <-Map 24 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_231] - PartitionCols:_col0, _col2 - Select Operator [SEL_230] (rows=144002668 width=135) - Output:["_col0","_col1","_col2"] - Filter Operator [FIL_229] (rows=144002668 width=135) - predicate:(ws_item_sk is not null and ws_order_number is not null and ws_web_site_sk is not null) - TableScan [TS_54] (rows=144002668 width=135) - default@web_sales,web_sales,Tbl:COMPLETE,Col:NONE,Output:["ws_item_sk","ws_web_site_sk","ws_order_number"] - <-Map 26 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_234] + Merge Join Operator [MERGEJOIN_264] (rows=158402938 width=135) + Conds:RS_361._col0, _col2=RS_348._col1, _col2(Inner),Output:["_col1","_col3","_col6","_col7"] + <-Map 31 [SIMPLE_EDGE] vectorized + PARTITION_ONLY_SHUFFLE [RS_348] PartitionCols:_col1, _col2 - Select Operator [SEL_233] (rows=14398467 width=92) + Select Operator [SEL_347] (rows=14398467 width=92) Output:["_col0","_col1","_col2","_col3","_col4"] - Filter Operator [FIL_232] (rows=14398467 width=92) + Filter Operator [FIL_346] (rows=14398467 width=92) predicate:(wr_item_sk is not null and wr_order_number is not null and wr_returned_date_sk is not null) TableScan [TS_57] (rows=14398467 width=92) default@web_returns,web_returns,Tbl:COMPLETE,Col:NONE,Output:["wr_returned_date_sk","wr_item_sk","wr_order_number","wr_return_amt","wr_net_loss"] + <-Map 29 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_361] + PartitionCols:_col0, _col2 + Select Operator [SEL_360] (rows=144002668 width=135) + Output:["_col0","_col1","_col2"] + Filter Operator [FIL_359] (rows=144002668 width=135) + predicate:((ws_item_sk BETWEEN DynamicValue(RS_61_web_returns_wr_item_sk_min) AND DynamicValue(RS_61_web_returns_wr_item_sk_max) and in_bloom_filter(ws_item_sk, DynamicValue(RS_61_web_returns_wr_item_sk_bloom_filter))) and (ws_order_number BETWEEN DynamicValue(RS_61_web_returns_wr_order_number_min) AND DynamicValue(RS_61_web_returns_wr_order_number_max) and in_bloom_filter(ws_order_number, DynamicValue(RS_61_web_returns_wr_order_number_bloom_filter))) and (ws_web_site_sk BETWEEN DynamicValue(RS_76_web_site_web_site_sk_min) AND DynamicValue(RS_76_web_site_web_site_sk_max) and in_bloom_filter(ws_web_site_sk, DynamicValue(RS_76_web_site_web_site_sk_bloom_filter))) and ws_item_sk is not null and ws_order_number is not null and ws_web_site_sk is not null) + TableScan [TS_54] (rows=144002668 width=135) + default@web_sales,web_sales,Tbl:COMPLETE,Col:NONE,Output:["ws_item_sk","ws_web_site_sk","ws_order_number"] + <-Reducer 35 [BROADCAST_EDGE] vectorized + BROADCAST [RS_342] + Please refer to the previous Group By Operator [GBY_340] + <-Reducer 32 [BROADCAST_EDGE] vectorized + BROADCAST [RS_356] + Group By Operator [GBY_355] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=14398467)"] + <-Map 31 [CUSTOM_SIMPLE_EDGE] vectorized + PARTITION_ONLY_SHUFFLE [RS_353] + Group By Operator [GBY_351] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=14398467)"] + Select Operator [SEL_349] (rows=14398467 width=92) + Output:["_col0"] + Please refer to the previous Select Operator [SEL_347] + <-Reducer 33 [BROADCAST_EDGE] vectorized + BROADCAST [RS_358] + Group By Operator [GBY_357] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=14398467)"] + <-Map 31 [CUSTOM_SIMPLE_EDGE] vectorized + PARTITION_ONLY_SHUFFLE [RS_354] + Group By Operator [GBY_352] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=14398467)"] + Select Operator [SEL_350] (rows=14398467 width=92) + Output:["_col0"] + Please refer to the previous Select Operator [SEL_347] <-Reducer 5 [CONTAINS] vectorized - Reduce Output Operator [RS_196] + Reduce Output Operator [RS_298] PartitionCols:_col0, _col1, _col2 - Group By Operator [GBY_195] (rows=2273797803 width=110) + Group By Operator [GBY_297] (rows=2273797803 width=110) Output:["_col0","_col1","_col2","_col3","_col4","_col5"],aggregations:["sum(_col2)","sum(_col3)","sum(_col4)"],keys:_col0, _col1, 0L - Select Operator [SEL_194] (rows=383320021 width=87) + Select Operator [SEL_296] (rows=383320021 width=87) Output:["_col0","_col1","_col2","_col3","_col4"] - Group By Operator [GBY_193] (rows=383320021 width=87) + Group By Operator [GBY_295] (rows=383320021 width=87) Output:["_col0","_col1","_col2","_col3","_col4"],aggregations:["sum(VALUE._col0)","sum(VALUE._col1)","sum(VALUE._col2)","sum(VALUE._col3)"],keys:KEY._col0 <-Reducer 4 [SIMPLE_EDGE] SHUFFLE [RS_22] PartitionCols:_col0 Group By Operator [GBY_21] (rows=766640042 width=87) Output:["_col0","_col1","_col2","_col3","_col4"],aggregations:["sum(_col2)","sum(_col4)","sum(_col3)","sum(_col5)"],keys:_col9 - Merge Join Operator [MERGEJOIN_133] (rows=766640042 width=87) - Conds:RS_17._col0=RS_192._col0(Inner),Output:["_col2","_col3","_col4","_col5","_col9"] - <-Map 17 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_192] + Merge Join Operator [MERGEJOIN_219] (rows=766640042 width=87) + Conds:RS_17._col0=RS_286._col0(Inner),Output:["_col2","_col3","_col4","_col5","_col9"] + <-Map 20 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_286] PartitionCols:_col0 - Select Operator [SEL_191] (rows=1704 width=1910) + Select Operator [SEL_285] (rows=1704 width=1910) Output:["_col0","_col1"] - Filter Operator [FIL_190] (rows=1704 width=1910) + Filter Operator [FIL_284] (rows=1704 width=1910) predicate:s_store_sk is not null TableScan [TS_11] (rows=1704 width=1910) default@store,store,Tbl:COMPLETE,Col:NONE,Output:["s_store_sk","s_store_id"] <-Reducer 3 [SIMPLE_EDGE] SHUFFLE [RS_17] PartitionCols:_col0 - Merge Join Operator [MERGEJOIN_132] (rows=696945478 width=87) - Conds:Union 2._col1=RS_187._col0(Inner),Output:["_col0","_col2","_col3","_col4","_col5"] + Merge Join Operator [MERGEJOIN_218] (rows=696945478 width=87) + Conds:Union 2._col1=RS_270._col0(Inner),Output:["_col0","_col2","_col3","_col4","_col5"] <-Map 10 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_187] + SHUFFLE [RS_270] PartitionCols:_col0 - Please refer to the previous Select Operator [SEL_186] + Please refer to the previous Select Operator [SEL_269] <-Union 2 [SIMPLE_EDGE] <-Map 1 [CONTAINS] vectorized - Reduce Output Operator [RS_184] + Reduce Output Operator [RS_294] PartitionCols:_col1 - Select Operator [SEL_183] (rows=575995635 width=88) + Select Operator [SEL_293] (rows=575995635 width=88) Output:["_col0","_col1","_col2","_col3","_col4","_col5"] - Filter Operator [FIL_182] (rows=575995635 width=88) - predicate:(ss_sold_date_sk is not null and ss_store_sk is not null) - TableScan [TS_138] (rows=575995635 width=88) + Filter Operator [FIL_292] (rows=575995635 width=88) + predicate:((ss_sold_date_sk BETWEEN DynamicValue(RS_15_date_dim_d_date_sk_min) AND DynamicValue(RS_15_date_dim_d_date_sk_max) and in_bloom_filter(ss_sold_date_sk, DynamicValue(RS_15_date_dim_d_date_sk_bloom_filter))) and (ss_store_sk BETWEEN DynamicValue(RS_18_store_s_store_sk_min) AND DynamicValue(RS_18_store_s_store_sk_max) and in_bloom_filter(ss_store_sk, DynamicValue(RS_18_store_s_store_sk_bloom_filter))) and ss_sold_date_sk is not null and ss_store_sk is not null) + TableScan [TS_224] (rows=575995635 width=88) Output:["ss_sold_date_sk","ss_store_sk","ss_ext_sales_price","ss_net_profit"] + <-Reducer 11 [BROADCAST_EDGE] vectorized + BROADCAST [RS_283] + Group By Operator [GBY_282] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] + <-Map 10 [CUSTOM_SIMPLE_EDGE] vectorized + SHUFFLE [RS_279] + Group By Operator [GBY_276] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] + Select Operator [SEL_271] (rows=8116 width=1119) + Output:["_col0"] + Please refer to the previous Select Operator [SEL_269] + <-Reducer 21 [BROADCAST_EDGE] vectorized + BROADCAST [RS_291] + Group By Operator [GBY_290] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] + <-Map 20 [CUSTOM_SIMPLE_EDGE] vectorized + SHUFFLE [RS_289] + Group By Operator [GBY_288] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] + Select Operator [SEL_287] (rows=1704 width=1910) + Output:["_col0"] + Please refer to the previous Select Operator [SEL_285] <-Map 9 [CONTAINS] vectorized - Reduce Output Operator [RS_205] + Reduce Output Operator [RS_307] PartitionCols:_col1 - Select Operator [SEL_204] (rows=57591150 width=77) + Select Operator [SEL_306] (rows=57591150 width=77) Output:["_col0","_col1","_col2","_col3","_col4","_col5"] - Filter Operator [FIL_203] (rows=57591150 width=77) + Filter Operator [FIL_305] (rows=57591150 width=77) predicate:(sr_returned_date_sk is not null and sr_store_sk is not null) - TableScan [TS_148] (rows=57591150 width=77) + TableScan [TS_234] (rows=57591150 width=77) Output:["sr_returned_date_sk","sr_store_sk","sr_return_amt","sr_net_loss"]
http://git-wip-us.apache.org/repos/asf/hive/blob/f2c4f319/ql/src/test/results/clientpositive/perf/tez/query50.q.out ---------------------------------------------------------------------- diff --git a/ql/src/test/results/clientpositive/perf/tez/query50.q.out b/ql/src/test/results/clientpositive/perf/tez/query50.q.out index 8ea0d5c..e723140 100644 --- a/ql/src/test/results/clientpositive/perf/tez/query50.q.out +++ b/ql/src/test/results/clientpositive/perf/tez/query50.q.out @@ -117,26 +117,32 @@ POSTHOOK: type: QUERY Plan optimized by CBO. Vertex dependency in root stage -Reducer 2 <- Map 1 (SIMPLE_EDGE), Map 8 (SIMPLE_EDGE) -Reducer 3 <- Map 9 (SIMPLE_EDGE), Reducer 2 (SIMPLE_EDGE) -Reducer 4 <- Map 10 (SIMPLE_EDGE), Reducer 3 (SIMPLE_EDGE) -Reducer 5 <- Map 11 (SIMPLE_EDGE), Reducer 4 (SIMPLE_EDGE) +Map 12 <- Reducer 10 (BROADCAST_EDGE), Reducer 14 (BROADCAST_EDGE), Reducer 16 (BROADCAST_EDGE), Reducer 8 (BROADCAST_EDGE), Reducer 9 (BROADCAST_EDGE) +Reducer 10 <- Reducer 2 (CUSTOM_SIMPLE_EDGE) +Reducer 14 <- Map 13 (CUSTOM_SIMPLE_EDGE) +Reducer 16 <- Map 15 (CUSTOM_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 13 (SIMPLE_EDGE), Reducer 3 (SIMPLE_EDGE) +Reducer 5 <- Map 15 (SIMPLE_EDGE), Reducer 4 (SIMPLE_EDGE) Reducer 6 <- Reducer 5 (SIMPLE_EDGE) Reducer 7 <- Reducer 6 (SIMPLE_EDGE) +Reducer 8 <- Reducer 2 (CUSTOM_SIMPLE_EDGE) +Reducer 9 <- Reducer 2 (CUSTOM_SIMPLE_EDGE) Stage-0 Fetch Operator limit:100 Stage-1 Reducer 7 vectorized - File Output Operator [FS_79] - Limit [LIM_78] (rows=100 width=88) + File Output Operator [FS_155] + Limit [LIM_154] (rows=100 width=88) Number of rows:100 - Select Operator [SEL_77] (rows=383325119 width=88) + Select Operator [SEL_153] (rows=383325119 width=88) Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6","_col7","_col8","_col9","_col10","_col11","_col12","_col13","_col14"] <-Reducer 6 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_76] - Group By Operator [GBY_75] (rows=383325119 width=88) + SHUFFLE [RS_152] + Group By Operator [GBY_151] (rows=383325119 width=88) Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6","_col7","_col8","_col9","_col10","_col11","_col12","_col13","_col14"],aggregations:["sum(VALUE._col0)","sum(VALUE._col1)","sum(VALUE._col2)","sum(VALUE._col3)","sum(VALUE._col4)"],keys:KEY._col0, KEY._col1, KEY._col2, KEY._col3, KEY._col4, KEY._col5, KEY._col6, KEY._col7, KEY._col8, KEY._col9 <-Reducer 5 [SIMPLE_EDGE] SHUFFLE [RS_30] @@ -145,66 +151,121 @@ Stage-0 Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6","_col7","_col8","_col9","_col10","_col11","_col12","_col13","_col14"],aggregations:["sum(_col10)","sum(_col11)","sum(_col12)","sum(_col13)","sum(_col14)"],keys:_col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8, _col9 Select Operator [SEL_27] (rows=766650239 width=88) Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6","_col7","_col8","_col9","_col10","_col11","_col12","_col13","_col14"] - Merge Join Operator [MERGEJOIN_59] (rows=766650239 width=88) - Conds:RS_24._col10=RS_74._col0(Inner),Output:["_col0","_col7","_col14","_col15","_col16","_col17","_col18","_col19","_col20","_col21","_col22","_col23"] - <-Map 11 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_74] + Merge Join Operator [MERGEJOIN_119] (rows=766650239 width=88) + Conds:RS_24._col10=RS_142._col0(Inner),Output:["_col0","_col7","_col14","_col15","_col16","_col17","_col18","_col19","_col20","_col21","_col22","_col23"] + <-Map 15 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_142] PartitionCols:_col0 - Select Operator [SEL_73] (rows=1704 width=1910) + Select Operator [SEL_141] (rows=1704 width=1910) Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6","_col7","_col8","_col9","_col10"] - Filter Operator [FIL_72] (rows=1704 width=1910) + Filter Operator [FIL_140] (rows=1704 width=1910) predicate:s_store_sk is not null TableScan [TS_12] (rows=1704 width=1910) default@store,store,Tbl:COMPLETE,Col:NONE,Output:["s_store_sk","s_store_name","s_company_id","s_street_number","s_street_name","s_street_type","s_suite_number","s_city","s_county","s_state","s_zip"] <-Reducer 4 [SIMPLE_EDGE] SHUFFLE [RS_24] PartitionCols:_col10 - Merge Join Operator [MERGEJOIN_58] (rows=696954748 width=88) - Conds:RS_21._col7=RS_71._col0(Inner),Output:["_col0","_col7","_col10"] - <-Map 10 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_71] + Merge Join Operator [MERGEJOIN_118] (rows=696954748 width=88) + Conds:RS_21._col7=RS_134._col0(Inner),Output:["_col0","_col7","_col10"] + <-Map 13 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_134] PartitionCols:_col0 - Select Operator [SEL_70] (rows=73049 width=1119) + Select Operator [SEL_133] (rows=73049 width=1119) Output:["_col0"] - Filter Operator [FIL_69] (rows=73049 width=1119) + Filter Operator [FIL_132] (rows=73049 width=1119) predicate:d_date_sk is not null TableScan [TS_9] (rows=73049 width=1119) default@date_dim,d1,Tbl:COMPLETE,Col:NONE,Output:["d_date_sk"] <-Reducer 3 [SIMPLE_EDGE] SHUFFLE [RS_21] PartitionCols:_col7 - Merge Join Operator [MERGEJOIN_57] (rows=633595212 width=88) - Conds:RS_18._col1, _col2, _col3=RS_68._col1, _col2, _col4(Inner),Output:["_col0","_col7","_col10"] - <-Map 9 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_68] - PartitionCols:_col1, _col2, _col4 - Select Operator [SEL_67] (rows=575995635 width=88) - Output:["_col0","_col1","_col2","_col3","_col4"] - Filter Operator [FIL_66] (rows=575995635 width=88) - predicate:(ss_customer_sk is not null and ss_item_sk is not null and ss_sold_date_sk is not null and ss_store_sk is not null and ss_ticket_number is not null) - TableScan [TS_6] (rows=575995635 width=88) - default@store_sales,store_sales,Tbl:COMPLETE,Col:NONE,Output:["ss_sold_date_sk","ss_item_sk","ss_customer_sk","ss_store_sk","ss_ticket_number"] + Merge Join Operator [MERGEJOIN_117] (rows=633595212 width=88) + Conds:RS_18._col1, _col2, _col3=RS_150._col1, _col2, _col4(Inner),Output:["_col0","_col7","_col10"] <-Reducer 2 [SIMPLE_EDGE] - SHUFFLE [RS_18] + PARTITION_ONLY_SHUFFLE [RS_18] PartitionCols:_col1, _col2, _col3 - Merge Join Operator [MERGEJOIN_56] (rows=63350266 width=77) - Conds:RS_62._col0=RS_65._col0(Inner),Output:["_col0","_col1","_col2","_col3"] + Merge Join Operator [MERGEJOIN_116] (rows=63350266 width=77) + Conds:RS_122._col0=RS_125._col0(Inner),Output:["_col0","_col1","_col2","_col3"] <-Map 1 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_62] + SHUFFLE [RS_122] PartitionCols:_col0 - Select Operator [SEL_61] (rows=57591150 width=77) + Select Operator [SEL_121] (rows=57591150 width=77) Output:["_col0","_col1","_col2","_col3"] - Filter Operator [FIL_60] (rows=57591150 width=77) + Filter Operator [FIL_120] (rows=57591150 width=77) predicate:(sr_customer_sk is not null and sr_item_sk is not null and sr_returned_date_sk is not null and sr_ticket_number is not null) TableScan [TS_0] (rows=57591150 width=77) default@store_returns,store_returns,Tbl:COMPLETE,Col:NONE,Output:["sr_returned_date_sk","sr_item_sk","sr_customer_sk","sr_ticket_number"] - <-Map 8 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_65] + <-Map 11 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_125] PartitionCols:_col0 - Select Operator [SEL_64] (rows=18262 width=1119) + Select Operator [SEL_124] (rows=18262 width=1119) Output:["_col0"] - Filter Operator [FIL_63] (rows=18262 width=1119) + Filter Operator [FIL_123] (rows=18262 width=1119) predicate:((d_moy = 9) and (d_year = 2000) and d_date_sk is not null) TableScan [TS_3] (rows=73049 width=1119) default@date_dim,d2,Tbl:COMPLETE,Col:NONE,Output:["d_date_sk","d_year","d_moy"] + <-Map 12 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_150] + PartitionCols:_col1, _col2, _col4 + Select Operator [SEL_149] (rows=575995635 width=88) + Output:["_col0","_col1","_col2","_col3","_col4"] + Filter Operator [FIL_148] (rows=575995635 width=88) + predicate:((ss_customer_sk BETWEEN DynamicValue(RS_18_store_returns_sr_customer_sk_min) AND DynamicValue(RS_18_store_returns_sr_customer_sk_max) and in_bloom_filter(ss_customer_sk, DynamicValue(RS_18_store_returns_sr_customer_sk_bloom_filter))) and (ss_item_sk BETWEEN DynamicValue(RS_18_store_returns_sr_item_sk_min) AND DynamicValue(RS_18_store_returns_sr_item_sk_max) and in_bloom_filter(ss_item_sk, DynamicValue(RS_18_store_returns_sr_item_sk_bloom_filter))) and (ss_sold_date_sk BETWEEN DynamicValue(RS_22_d1_d_date_sk_min) AND DynamicValue(RS_22_d1_d_date_sk_max) and in_bloom_filter(ss_sold_date_sk, DynamicValue(RS_22_d1_d_date_sk_bloom_filter))) and (ss_store_sk BETWEEN DynamicValue(RS_25_store_s_store_sk_min) AND DynamicValue(RS_25_store_s_store_sk_max) and in_bloom_filter(ss_store_sk, DynamicValue(RS_25_store_s_store_sk_bloom_filter))) and (ss_ticket_number BETWEEN DynamicValue(RS_18_store_returns_sr_ticket_number_min) AND DynamicValue(RS_18_ store_returns_sr_ticket_number_max) and in_bloom_filter(ss_ticket_number, DynamicValue(RS_18_store_returns_sr_ticket_number_bloom_filter))) and ss_customer_sk is not null and ss_item_sk is not null and ss_sold_date_sk is not null and ss_store_sk is not null and ss_ticket_number is not null) + TableScan [TS_6] (rows=575995635 width=88) + default@store_sales,store_sales,Tbl:COMPLETE,Col:NONE,Output:["ss_sold_date_sk","ss_item_sk","ss_customer_sk","ss_store_sk","ss_ticket_number"] + <-Reducer 10 [BROADCAST_EDGE] vectorized + BROADCAST [RS_131] + Group By Operator [GBY_130] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=63350264)"] + <-Reducer 2 [CUSTOM_SIMPLE_EDGE] + PARTITION_ONLY_SHUFFLE [RS_93] + Group By Operator [GBY_92] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=63350264)"] + Select Operator [SEL_91] (rows=63350266 width=77) + Output:["_col0"] + Please refer to the previous Merge Join Operator [MERGEJOIN_116] + <-Reducer 14 [BROADCAST_EDGE] vectorized + BROADCAST [RS_139] + Group By Operator [GBY_138] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] + <-Map 13 [CUSTOM_SIMPLE_EDGE] vectorized + SHUFFLE [RS_137] + Group By Operator [GBY_136] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] + Select Operator [SEL_135] (rows=73049 width=1119) + Output:["_col0"] + Please refer to the previous Select Operator [SEL_133] + <-Reducer 16 [BROADCAST_EDGE] vectorized + BROADCAST [RS_147] + Group By Operator [GBY_146] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] + <-Map 15 [CUSTOM_SIMPLE_EDGE] vectorized + SHUFFLE [RS_145] + Group By Operator [GBY_144] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] + Select Operator [SEL_143] (rows=1704 width=1910) + Output:["_col0"] + Please refer to the previous Select Operator [SEL_141] + <-Reducer 8 [BROADCAST_EDGE] vectorized + BROADCAST [RS_127] + Group By Operator [GBY_126] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=63350264)"] + <-Reducer 2 [CUSTOM_SIMPLE_EDGE] + PARTITION_ONLY_SHUFFLE [RS_83] + Group By Operator [GBY_82] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=63350264)"] + Select Operator [SEL_81] (rows=63350266 width=77) + Output:["_col0"] + Please refer to the previous Merge Join Operator [MERGEJOIN_116] + <-Reducer 9 [BROADCAST_EDGE] vectorized + BROADCAST [RS_129] + Group By Operator [GBY_128] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=63350264)"] + <-Reducer 2 [CUSTOM_SIMPLE_EDGE] + PARTITION_ONLY_SHUFFLE [RS_88] + Group By Operator [GBY_87] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=63350264)"] + Select Operator [SEL_86] (rows=63350266 width=77) + Output:["_col0"] + Please refer to the previous Merge Join Operator [MERGEJOIN_116] http://git-wip-us.apache.org/repos/asf/hive/blob/f2c4f319/ql/src/test/results/clientpositive/perf/tez/query51.q.out ---------------------------------------------------------------------- diff --git a/ql/src/test/results/clientpositive/perf/tez/query51.q.out b/ql/src/test/results/clientpositive/perf/tez/query51.q.out index 8a0802e..be123ae 100644 --- a/ql/src/test/results/clientpositive/perf/tez/query51.q.out +++ b/ql/src/test/results/clientpositive/perf/tez/query51.q.out @@ -89,23 +89,27 @@ POSTHOOK: type: QUERY Plan optimized by CBO. Vertex dependency in root stage +Map 1 <- Reducer 8 (BROADCAST_EDGE) +Map 12 <- Reducer 11 (BROADCAST_EDGE) +Reducer 10 <- Reducer 9 (SIMPLE_EDGE) +Reducer 11 <- Map 7 (CUSTOM_SIMPLE_EDGE) Reducer 2 <- Map 1 (SIMPLE_EDGE), Map 7 (SIMPLE_EDGE) Reducer 3 <- Reducer 2 (SIMPLE_EDGE) -Reducer 4 <- Reducer 3 (SIMPLE_EDGE), Reducer 9 (SIMPLE_EDGE) +Reducer 4 <- Reducer 10 (SIMPLE_EDGE), Reducer 3 (SIMPLE_EDGE) Reducer 5 <- Reducer 4 (SIMPLE_EDGE) Reducer 6 <- Reducer 5 (SIMPLE_EDGE) -Reducer 8 <- Map 10 (SIMPLE_EDGE), Map 7 (SIMPLE_EDGE) -Reducer 9 <- Reducer 8 (SIMPLE_EDGE) +Reducer 8 <- Map 7 (CUSTOM_SIMPLE_EDGE) +Reducer 9 <- Map 12 (SIMPLE_EDGE), Map 7 (SIMPLE_EDGE) Stage-0 Fetch Operator limit:100 Stage-1 Reducer 6 vectorized - File Output Operator [FS_83] - Limit [LIM_82] (rows=100 width=88) + File Output Operator [FS_113] + Limit [LIM_112] (rows=100 width=88) Number of rows:100 - Select Operator [SEL_81] (rows=116159124 width=88) + Select Operator [SEL_111] (rows=116159124 width=88) Output:["_col0","_col1","_col2","_col3","_col4","_col5"] <-Reducer 5 [SIMPLE_EDGE] SHUFFLE [RS_50] @@ -120,69 +124,91 @@ Stage-0 <-Reducer 4 [SIMPLE_EDGE] SHUFFLE [RS_43] PartitionCols:CASE WHEN (_col3 is not null) THEN (_col3) ELSE (_col0) END - Merge Join Operator [MERGEJOIN_67] (rows=348477374 width=88) + Merge Join Operator [MERGEJOIN_87] (rows=348477374 width=88) Conds:RS_40._col0, _col1=RS_41._col0, _col1(Outer),Output:["_col0","_col1","_col2","_col3","_col4","_col5"] - <-Reducer 3 [SIMPLE_EDGE] - SHUFFLE [RS_40] + <-Reducer 10 [SIMPLE_EDGE] + SHUFFLE [RS_41] PartitionCols:_col0, _col1 - Select Operator [SEL_17] (rows=316797606 width=88) + Select Operator [SEL_37] (rows=79201469 width=135) Output:["_col0","_col1","_col2"] - PTF Operator [PTF_16] (rows=316797606 width=88) + PTF Operator [PTF_36] (rows=79201469 width=135) Function definitions:[{},{"name:":"windowingtablefunction","order by:":"_col1 ASC NULLS FIRST","partition by:":"_col0"}] - Group By Operator [GBY_12] (rows=316797606 width=88) + Group By Operator [GBY_32] (rows=79201469 width=135) Output:["_col0","_col1","_col2"],aggregations:["sum(VALUE._col0)"],keys:KEY._col0, KEY._col1 - <-Reducer 2 [SIMPLE_EDGE] - SHUFFLE [RS_11] + <-Reducer 9 [SIMPLE_EDGE] + SHUFFLE [RS_31] PartitionCols:_col0 - Group By Operator [GBY_10] (rows=633595212 width=88) + Group By Operator [GBY_30] (rows=158402938 width=135) Output:["_col0","_col1","_col2"],aggregations:["sum(_col2)"],keys:_col1, _col4 - Merge Join Operator [MERGEJOIN_65] (rows=633595212 width=88) - Conds:RS_70._col0=RS_73._col0(Inner),Output:["_col1","_col2","_col4"] + Merge Join Operator [MERGEJOIN_86] (rows=158402938 width=135) + Conds:RS_108._col0=RS_92._col0(Inner),Output:["_col1","_col2","_col4"] <-Map 7 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_73] + SHUFFLE [RS_92] PartitionCols:_col0 - Select Operator [SEL_72] (rows=8116 width=1119) + Select Operator [SEL_89] (rows=8116 width=1119) Output:["_col0","_col1"] - Filter Operator [FIL_71] (rows=8116 width=1119) + Filter Operator [FIL_88] (rows=8116 width=1119) predicate:(d_date_sk is not null and d_month_seq BETWEEN 1212 AND 1223) TableScan [TS_3] (rows=73049 width=1119) default@date_dim,date_dim,Tbl:COMPLETE,Col:NONE,Output:["d_date_sk","d_date","d_month_seq"] - <-Map 1 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_70] + <-Map 12 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_108] PartitionCols:_col0 - Select Operator [SEL_69] (rows=575995635 width=88) + Select Operator [SEL_107] (rows=144002668 width=135) Output:["_col0","_col1","_col2"] - Filter Operator [FIL_68] (rows=575995635 width=88) - predicate:(ss_item_sk is not null and ss_sold_date_sk is not null) - TableScan [TS_0] (rows=575995635 width=88) - default@store_sales,store_sales,Tbl:COMPLETE,Col:NONE,Output:["ss_sold_date_sk","ss_item_sk","ss_sales_price"] - <-Reducer 9 [SIMPLE_EDGE] - SHUFFLE [RS_41] + Filter Operator [FIL_106] (rows=144002668 width=135) + predicate:((ws_sold_date_sk BETWEEN DynamicValue(RS_27_date_dim_d_date_sk_min) AND DynamicValue(RS_27_date_dim_d_date_sk_max) and in_bloom_filter(ws_sold_date_sk, DynamicValue(RS_27_date_dim_d_date_sk_bloom_filter))) and ws_item_sk is not null and ws_sold_date_sk is not null) + TableScan [TS_20] (rows=144002668 width=135) + default@web_sales,web_sales,Tbl:COMPLETE,Col:NONE,Output:["ws_sold_date_sk","ws_item_sk","ws_sales_price"] + <-Reducer 11 [BROADCAST_EDGE] vectorized + BROADCAST [RS_105] + Group By Operator [GBY_104] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] + <-Map 7 [CUSTOM_SIMPLE_EDGE] vectorized + SHUFFLE [RS_97] + Group By Operator [GBY_95] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] + Select Operator [SEL_93] (rows=8116 width=1119) + Output:["_col0"] + Please refer to the previous Select Operator [SEL_89] + <-Reducer 3 [SIMPLE_EDGE] + SHUFFLE [RS_40] PartitionCols:_col0, _col1 - Select Operator [SEL_37] (rows=79201469 width=135) + Select Operator [SEL_17] (rows=316797606 width=88) Output:["_col0","_col1","_col2"] - PTF Operator [PTF_36] (rows=79201469 width=135) + PTF Operator [PTF_16] (rows=316797606 width=88) Function definitions:[{},{"name:":"windowingtablefunction","order by:":"_col1 ASC NULLS FIRST","partition by:":"_col0"}] - Group By Operator [GBY_32] (rows=79201469 width=135) + Group By Operator [GBY_12] (rows=316797606 width=88) Output:["_col0","_col1","_col2"],aggregations:["sum(VALUE._col0)"],keys:KEY._col0, KEY._col1 - <-Reducer 8 [SIMPLE_EDGE] - SHUFFLE [RS_31] + <-Reducer 2 [SIMPLE_EDGE] + SHUFFLE [RS_11] PartitionCols:_col0 - Group By Operator [GBY_30] (rows=158402938 width=135) + Group By Operator [GBY_10] (rows=633595212 width=88) Output:["_col0","_col1","_col2"],aggregations:["sum(_col2)"],keys:_col1, _col4 - Merge Join Operator [MERGEJOIN_66] (rows=158402938 width=135) - Conds:RS_78._col0=RS_74._col0(Inner),Output:["_col1","_col2","_col4"] + Merge Join Operator [MERGEJOIN_85] (rows=633595212 width=88) + Conds:RS_102._col0=RS_90._col0(Inner),Output:["_col1","_col2","_col4"] <-Map 7 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_74] + SHUFFLE [RS_90] PartitionCols:_col0 - Please refer to the previous Select Operator [SEL_72] - <-Map 10 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_78] + Please refer to the previous Select Operator [SEL_89] + <-Map 1 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_102] PartitionCols:_col0 - Select Operator [SEL_77] (rows=144002668 width=135) + Select Operator [SEL_101] (rows=575995635 width=88) Output:["_col0","_col1","_col2"] - Filter Operator [FIL_76] (rows=144002668 width=135) - predicate:(ws_item_sk is not null and ws_sold_date_sk is not null) - TableScan [TS_20] (rows=144002668 width=135) - default@web_sales,web_sales,Tbl:COMPLETE,Col:NONE,Output:["ws_sold_date_sk","ws_item_sk","ws_sales_price"] + Filter Operator [FIL_100] (rows=575995635 width=88) + predicate:((ss_sold_date_sk BETWEEN DynamicValue(RS_7_date_dim_d_date_sk_min) AND DynamicValue(RS_7_date_dim_d_date_sk_max) and in_bloom_filter(ss_sold_date_sk, DynamicValue(RS_7_date_dim_d_date_sk_bloom_filter))) and ss_item_sk is not null and ss_sold_date_sk is not null) + TableScan [TS_0] (rows=575995635 width=88) + default@store_sales,store_sales,Tbl:COMPLETE,Col:NONE,Output:["ss_sold_date_sk","ss_item_sk","ss_sales_price"] + <-Reducer 8 [BROADCAST_EDGE] vectorized + BROADCAST [RS_99] + Group By Operator [GBY_98] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] + <-Map 7 [CUSTOM_SIMPLE_EDGE] vectorized + SHUFFLE [RS_96] + Group By Operator [GBY_94] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] + Select Operator [SEL_91] (rows=8116 width=1119) + Output:["_col0"] + Please refer to the previous Select Operator [SEL_89] http://git-wip-us.apache.org/repos/asf/hive/blob/f2c4f319/ql/src/test/results/clientpositive/perf/tez/query52.q.out ---------------------------------------------------------------------- diff --git a/ql/src/test/results/clientpositive/perf/tez/query52.q.out b/ql/src/test/results/clientpositive/perf/tez/query52.q.out index f3d03c2..eec20fb 100644 --- a/ql/src/test/results/clientpositive/perf/tez/query52.q.out +++ b/ql/src/test/results/clientpositive/perf/tez/query52.q.out @@ -43,64 +43,89 @@ POSTHOOK: type: QUERY Plan optimized by CBO. Vertex dependency in root stage +Map 1 <- Reducer 7 (BROADCAST_EDGE), Reducer 9 (BROADCAST_EDGE) Reducer 2 <- Map 1 (SIMPLE_EDGE), Map 6 (SIMPLE_EDGE) -Reducer 3 <- Map 7 (SIMPLE_EDGE), Reducer 2 (SIMPLE_EDGE) +Reducer 3 <- Map 8 (SIMPLE_EDGE), Reducer 2 (SIMPLE_EDGE) Reducer 4 <- Reducer 3 (SIMPLE_EDGE) Reducer 5 <- Reducer 4 (SIMPLE_EDGE) +Reducer 7 <- Map 6 (CUSTOM_SIMPLE_EDGE) +Reducer 9 <- Map 8 (CUSTOM_SIMPLE_EDGE) Stage-0 Fetch Operator limit:-1 Stage-1 Reducer 5 vectorized - File Output Operator [FS_49] - Select Operator [SEL_48] (rows=100 width=88) + File Output Operator [FS_79] + Select Operator [SEL_78] (rows=100 width=88) Output:["_col0","_col1","_col2","_col3"] - Limit [LIM_47] (rows=100 width=88) + Limit [LIM_77] (rows=100 width=88) Number of rows:100 - Select Operator [SEL_46] (rows=348477374 width=88) + Select Operator [SEL_76] (rows=348477374 width=88) Output:["_col0","_col1","_col2"] <-Reducer 4 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_45] - Group By Operator [GBY_44] (rows=348477374 width=88) + SHUFFLE [RS_75] + Group By Operator [GBY_74] (rows=348477374 width=88) Output:["_col0","_col1","_col2"],aggregations:["sum(VALUE._col0)"],keys:KEY._col0, KEY._col1 <-Reducer 3 [SIMPLE_EDGE] SHUFFLE [RS_17] PartitionCols:_col0, _col1 Group By Operator [GBY_16] (rows=696954748 width=88) Output:["_col0","_col1","_col2"],aggregations:["sum(_col2)"],keys:_col7, _col8 - Merge Join Operator [MERGEJOIN_34] (rows=696954748 width=88) - Conds:RS_12._col1=RS_43._col0(Inner),Output:["_col2","_col7","_col8"] - <-Map 7 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_43] + Merge Join Operator [MERGEJOIN_54] (rows=696954748 width=88) + Conds:RS_12._col1=RS_65._col0(Inner),Output:["_col2","_col7","_col8"] + <-Map 8 [SIMPLE_EDGE] vectorized + PARTITION_ONLY_SHUFFLE [RS_65] PartitionCols:_col0 - Select Operator [SEL_42] (rows=231000 width=1436) + Select Operator [SEL_64] (rows=231000 width=1436) Output:["_col0","_col1","_col2"] - Filter Operator [FIL_41] (rows=231000 width=1436) + Filter Operator [FIL_63] (rows=231000 width=1436) predicate:((i_manager_id = 1) and i_item_sk is not null) TableScan [TS_6] (rows=462000 width=1436) default@item,item,Tbl:COMPLETE,Col:NONE,Output:["i_item_sk","i_brand_id","i_brand","i_manager_id"] <-Reducer 2 [SIMPLE_EDGE] SHUFFLE [RS_12] PartitionCols:_col1 - Merge Join Operator [MERGEJOIN_33] (rows=633595212 width=88) - Conds:RS_37._col0=RS_40._col0(Inner),Output:["_col1","_col2"] - <-Map 1 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_37] - PartitionCols:_col0 - Select Operator [SEL_36] (rows=575995635 width=88) - Output:["_col0","_col1","_col2"] - Filter Operator [FIL_35] (rows=575995635 width=88) - predicate:(ss_item_sk is not null and ss_sold_date_sk is not null) - TableScan [TS_0] (rows=575995635 width=88) - default@store_sales,store_sales,Tbl:COMPLETE,Col:NONE,Output:["ss_sold_date_sk","ss_item_sk","ss_ext_sales_price"] + Merge Join Operator [MERGEJOIN_53] (rows=633595212 width=88) + Conds:RS_73._col0=RS_57._col0(Inner),Output:["_col1","_col2"] <-Map 6 [SIMPLE_EDGE] vectorized - SHUFFLE [RS_40] + PARTITION_ONLY_SHUFFLE [RS_57] PartitionCols:_col0 - Select Operator [SEL_39] (rows=18262 width=1119) + Select Operator [SEL_56] (rows=18262 width=1119) Output:["_col0"] - Filter Operator [FIL_38] (rows=18262 width=1119) + Filter Operator [FIL_55] (rows=18262 width=1119) predicate:((d_moy = 12) and (d_year = 1998) and d_date_sk is not null) TableScan [TS_3] (rows=73049 width=1119) default@date_dim,dt,Tbl:COMPLETE,Col:NONE,Output:["d_date_sk","d_year","d_moy"] + <-Map 1 [SIMPLE_EDGE] vectorized + SHUFFLE [RS_73] + PartitionCols:_col0 + Select Operator [SEL_72] (rows=575995635 width=88) + Output:["_col0","_col1","_col2"] + Filter Operator [FIL_71] (rows=575995635 width=88) + predicate:((ss_item_sk BETWEEN DynamicValue(RS_13_item_i_item_sk_min) AND DynamicValue(RS_13_item_i_item_sk_max) and in_bloom_filter(ss_item_sk, DynamicValue(RS_13_item_i_item_sk_bloom_filter))) and (ss_sold_date_sk BETWEEN DynamicValue(RS_10_dt_d_date_sk_min) AND DynamicValue(RS_10_dt_d_date_sk_max) and in_bloom_filter(ss_sold_date_sk, DynamicValue(RS_10_dt_d_date_sk_bloom_filter))) and ss_item_sk is not null and ss_sold_date_sk is not null) + TableScan [TS_0] (rows=575995635 width=88) + default@store_sales,store_sales,Tbl:COMPLETE,Col:NONE,Output:["ss_sold_date_sk","ss_item_sk","ss_ext_sales_price"] + <-Reducer 7 [BROADCAST_EDGE] vectorized + BROADCAST [RS_62] + Group By Operator [GBY_61] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(VALUE._col0)","max(VALUE._col1)","bloom_filter(VALUE._col2, expectedEntries=1000000)"] + <-Map 6 [CUSTOM_SIMPLE_EDGE] vectorized + PARTITION_ONLY_SHUFFLE [RS_60] + Group By Operator [GBY_59] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] + Select Operator [SEL_58] (rows=18262 width=1119) + Output:["_col0"] + Please refer to the previous Select Operator [SEL_56] + <-Reducer 9 [BROADCAST_EDGE] vectorized + BROADCAST [RS_70] + Group By Operator [GBY_69] (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_68] + Group By Operator [GBY_67] (rows=1 width=12) + Output:["_col0","_col1","_col2"],aggregations:["min(_col0)","max(_col0)","bloom_filter(_col0, expectedEntries=1000000)"] + Select Operator [SEL_66] (rows=231000 width=1436) + Output:["_col0"] + Please refer to the previous Select Operator [SEL_64]
