You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@hive.apache.org by jc...@apache.org on 2018/07/19 21:24:15 UTC
[02/13] hive git commit: HIVE-17896: TopNKey: Create a standalone
vectorizable TopNKey operator (Teddy Choi,
reviewed by Jesus Camacho Rodriguez)
http://git-wip-us.apache.org/repos/asf/hive/blob/cc294d32/ql/src/test/results/clientpositive/perf/tez/query80.q.out
----------------------------------------------------------------------
diff --git a/ql/src/test/results/clientpositive/perf/tez/query80.q.out b/ql/src/test/results/clientpositive/perf/tez/query80.q.out
index fb4cd56..f250c5a 100644
--- a/ql/src/test/results/clientpositive/perf/tez/query80.q.out
+++ b/ql/src/test/results/clientpositive/perf/tez/query80.q.out
@@ -217,262 +217,268 @@ Stage-0
limit:100
Stage-1
Reducer 10 vectorized
- File Output Operator [FS_266]
- Limit [LIM_265] (rows=100 width=108)
+ File Output Operator [FS_271]
+ Limit [LIM_270] (rows=100 width=108)
Number of rows:100
- Select Operator [SEL_264] (rows=1217531358 width=108)
+ Select Operator [SEL_269] (rows=1217531358 width=108)
Output:["_col0","_col1","_col2","_col3","_col4"]
<-Reducer 9 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_263]
- Select Operator [SEL_262] (rows=1217531358 width=108)
+ SHUFFLE [RS_268]
+ Select Operator [SEL_267] (rows=1217531358 width=108)
Output:["_col0","_col1","_col2","_col3","_col4"]
- Group By Operator [GBY_261] (rows=1217531358 width=108)
+ Group By Operator [GBY_266] (rows=1217531358 width=108)
Output:["_col0","_col1","_col3","_col4","_col5"],aggregations:["sum(VALUE._col0)","sum(VALUE._col1)","sum(VALUE._col2)"],keys:KEY._col0, KEY._col1, KEY._col2
<-Union 8 [SIMPLE_EDGE]
<-Reducer 17 [CONTAINS] vectorized
- Reduce Output Operator [RS_279]
+ Reduce Output Operator [RS_285]
PartitionCols:_col0, _col1, _col2
- Group By Operator [GBY_278] (rows=2435062716 width=108)
+ Group By Operator [GBY_284] (rows=2435062716 width=108)
Output:["_col0","_col1","_col2","_col3","_col4","_col5"],aggregations:["sum(_col2)","sum(_col3)","sum(_col4)"],keys:_col0, _col1, 0L
- Select Operator [SEL_277] (rows=231905279 width=135)
- Output:["_col0","_col1","_col2","_col3","_col4"]
- Group By Operator [GBY_276] (rows=231905279 width=135)
- Output:["_col0","_col1","_col2","_col3"],aggregations:["sum(VALUE._col0)","sum(VALUE._col1)","sum(VALUE._col2)"],keys:KEY._col0
- <-Reducer 16 [SIMPLE_EDGE]
- SHUFFLE [RS_75]
- PartitionCols:_col0
- Group By Operator [GBY_74] (rows=463810558 width=135)
- Output:["_col0","_col1","_col2","_col3"],aggregations:["sum(_col1)","sum(_col2)","sum(_col3)"],keys:_col0
- Select Operator [SEL_72] (rows=463810558 width=135)
- Output:["_col0","_col1","_col2","_col3"]
- Merge Join Operator [MERGEJOIN_212] (rows=463810558 width=135)
- Conds:RS_69._col1=RS_275._col0(Inner),Output:["_col5","_col6","_col9","_col10","_col18"]
- <-Map 29 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_275]
- PartitionCols:_col0
- Select Operator [SEL_274] (rows=46000 width=460)
- Output:["_col0","_col1"]
- Filter Operator [FIL_273] (rows=46000 width=460)
- predicate:cp_catalog_page_sk is not null
- TableScan [TS_54] (rows=46000 width=460)
- default@catalog_page,catalog_page,Tbl:COMPLETE,Col:NONE,Output:["cp_catalog_page_sk","cp_catalog_page_id"]
- <-Reducer 15 [SIMPLE_EDGE]
- SHUFFLE [RS_69]
- PartitionCols:_col1
- Merge Join Operator [MERGEJOIN_211] (rows=421645953 width=135)
- Conds:RS_66._col3=RS_252._col0(Inner),Output:["_col1","_col5","_col6","_col9","_col10"]
- <-Map 24 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_252]
- PartitionCols:_col0
- Select Operator [SEL_250] (rows=1150 width=1179)
- Output:["_col0"]
- Filter Operator [FIL_249] (rows=1150 width=1179)
- predicate:((p_channel_tv = 'N') and p_promo_sk is not null)
- TableScan [TS_12] (rows=2300 width=1179)
- default@promotion,promotion,Tbl:COMPLETE,Col:NONE,Output:["p_promo_sk","p_channel_tv"]
- <-Reducer 14 [SIMPLE_EDGE]
- SHUFFLE [RS_66]
- PartitionCols:_col3
- Merge Join Operator [MERGEJOIN_210] (rows=383314495 width=135)
- Conds:RS_63._col2=RS_247._col0(Inner),Output:["_col1","_col3","_col5","_col6","_col9","_col10"]
- <-Map 23 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_247]
- PartitionCols:_col0
- Select Operator [SEL_245] (rows=154000 width=1436)
- Output:["_col0"]
- Filter Operator [FIL_244] (rows=154000 width=1436)
- predicate:((i_current_price > 50) and i_item_sk is not null)
- TableScan [TS_9] (rows=462000 width=1436)
- default@item,item,Tbl:COMPLETE,Col:NONE,Output:["i_item_sk","i_current_price"]
- <-Reducer 13 [SIMPLE_EDGE]
- SHUFFLE [RS_63]
- PartitionCols:_col2
- Merge Join Operator [MERGEJOIN_209] (rows=348467716 width=135)
- Conds:RS_60._col0=RS_242._col0(Inner),Output:["_col1","_col2","_col3","_col5","_col6","_col9","_col10"]
- <-Map 12 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_242]
- PartitionCols:_col0
- Select Operator [SEL_240] (rows=8116 width=1119)
- Output:["_col0"]
- Filter Operator [FIL_239] (rows=8116 width=1119)
- predicate:(CAST( d_date AS TIMESTAMP) BETWEEN TIMESTAMP'1998-08-04 00:00:00' AND TIMESTAMP'1998-09-03 00:00:00' and d_date_sk is not null)
- TableScan [TS_6] (rows=73049 width=1119)
- default@date_dim,date_dim,Tbl:COMPLETE,Col:NONE,Output:["d_date_sk","d_date"]
- <-Reducer 27 [SIMPLE_EDGE]
- SHUFFLE [RS_60]
- PartitionCols:_col0
- Merge Join Operator [MERGEJOIN_208] (rows=316788826 width=135)
- Conds:RS_269._col2, _col4=RS_272._col0, _col1(Left Outer),Output:["_col0","_col1","_col2","_col3","_col5","_col6","_col9","_col10"]
- <-Map 26 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_269]
- PartitionCols:_col2, _col4
- Select Operator [SEL_268] (rows=287989836 width=135)
- Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6"]
- Filter Operator [FIL_267] (rows=287989836 width=135)
- predicate:(cs_catalog_page_sk is not null and cs_item_sk is not null and cs_promo_sk is not null and cs_sold_date_sk is not null)
- TableScan [TS_39] (rows=287989836 width=135)
- default@catalog_sales,catalog_sales,Tbl:COMPLETE,Col:NONE,Output:["cs_sold_date_sk","cs_catalog_page_sk","cs_item_sk","cs_promo_sk","cs_order_number","cs_ext_sales_price","cs_net_profit"]
- <-Map 28 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_272]
- PartitionCols:_col0, _col1
- Select Operator [SEL_271] (rows=28798881 width=106)
- Output:["_col0","_col1","_col2","_col3"]
- Filter Operator [FIL_270] (rows=28798881 width=106)
- predicate:cr_item_sk is not null
- TableScan [TS_42] (rows=28798881 width=106)
- default@catalog_returns,catalog_returns,Tbl:COMPLETE,Col:NONE,Output:["cr_item_sk","cr_order_number","cr_return_amount","cr_net_loss"]
+ Top N Key Operator [TNK_283] (rows=811687572 width=108)
+ keys:_col0, _col1, 0L,sort order:+++,top n:100
+ Select Operator [SEL_282] (rows=231905279 width=135)
+ Output:["_col0","_col1","_col2","_col3","_col4"]
+ Group By Operator [GBY_281] (rows=231905279 width=135)
+ Output:["_col0","_col1","_col2","_col3"],aggregations:["sum(VALUE._col0)","sum(VALUE._col1)","sum(VALUE._col2)"],keys:KEY._col0
+ <-Reducer 16 [SIMPLE_EDGE]
+ SHUFFLE [RS_75]
+ PartitionCols:_col0
+ Group By Operator [GBY_74] (rows=463810558 width=135)
+ Output:["_col0","_col1","_col2","_col3"],aggregations:["sum(_col1)","sum(_col2)","sum(_col3)"],keys:_col0
+ Select Operator [SEL_72] (rows=463810558 width=135)
+ Output:["_col0","_col1","_col2","_col3"]
+ Merge Join Operator [MERGEJOIN_213] (rows=463810558 width=135)
+ Conds:RS_69._col1=RS_280._col0(Inner),Output:["_col5","_col6","_col9","_col10","_col18"]
+ <-Map 29 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_280]
+ PartitionCols:_col0
+ Select Operator [SEL_279] (rows=46000 width=460)
+ Output:["_col0","_col1"]
+ Filter Operator [FIL_278] (rows=46000 width=460)
+ predicate:cp_catalog_page_sk is not null
+ TableScan [TS_54] (rows=46000 width=460)
+ default@catalog_page,catalog_page,Tbl:COMPLETE,Col:NONE,Output:["cp_catalog_page_sk","cp_catalog_page_id"]
+ <-Reducer 15 [SIMPLE_EDGE]
+ SHUFFLE [RS_69]
+ PartitionCols:_col1
+ Merge Join Operator [MERGEJOIN_212] (rows=421645953 width=135)
+ Conds:RS_66._col3=RS_256._col0(Inner),Output:["_col1","_col5","_col6","_col9","_col10"]
+ <-Map 24 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_256]
+ PartitionCols:_col0
+ Select Operator [SEL_254] (rows=1150 width=1179)
+ Output:["_col0"]
+ Filter Operator [FIL_253] (rows=1150 width=1179)
+ predicate:((p_channel_tv = 'N') and p_promo_sk is not null)
+ TableScan [TS_12] (rows=2300 width=1179)
+ default@promotion,promotion,Tbl:COMPLETE,Col:NONE,Output:["p_promo_sk","p_channel_tv"]
+ <-Reducer 14 [SIMPLE_EDGE]
+ SHUFFLE [RS_66]
+ PartitionCols:_col3
+ Merge Join Operator [MERGEJOIN_211] (rows=383314495 width=135)
+ Conds:RS_63._col2=RS_251._col0(Inner),Output:["_col1","_col3","_col5","_col6","_col9","_col10"]
+ <-Map 23 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_251]
+ PartitionCols:_col0
+ Select Operator [SEL_249] (rows=154000 width=1436)
+ Output:["_col0"]
+ Filter Operator [FIL_248] (rows=154000 width=1436)
+ predicate:((i_current_price > 50) and i_item_sk is not null)
+ TableScan [TS_9] (rows=462000 width=1436)
+ default@item,item,Tbl:COMPLETE,Col:NONE,Output:["i_item_sk","i_current_price"]
+ <-Reducer 13 [SIMPLE_EDGE]
+ SHUFFLE [RS_63]
+ PartitionCols:_col2
+ Merge Join Operator [MERGEJOIN_210] (rows=348467716 width=135)
+ Conds:RS_60._col0=RS_246._col0(Inner),Output:["_col1","_col2","_col3","_col5","_col6","_col9","_col10"]
+ <-Map 12 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_246]
+ PartitionCols:_col0
+ Select Operator [SEL_244] (rows=8116 width=1119)
+ Output:["_col0"]
+ Filter Operator [FIL_243] (rows=8116 width=1119)
+ predicate:(CAST( d_date AS TIMESTAMP) BETWEEN TIMESTAMP'1998-08-04 00:00:00' AND TIMESTAMP'1998-09-03 00:00:00' and d_date_sk is not null)
+ TableScan [TS_6] (rows=73049 width=1119)
+ default@date_dim,date_dim,Tbl:COMPLETE,Col:NONE,Output:["d_date_sk","d_date"]
+ <-Reducer 27 [SIMPLE_EDGE]
+ SHUFFLE [RS_60]
+ PartitionCols:_col0
+ Merge Join Operator [MERGEJOIN_209] (rows=316788826 width=135)
+ Conds:RS_274._col2, _col4=RS_277._col0, _col1(Left Outer),Output:["_col0","_col1","_col2","_col3","_col5","_col6","_col9","_col10"]
+ <-Map 26 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_274]
+ PartitionCols:_col2, _col4
+ Select Operator [SEL_273] (rows=287989836 width=135)
+ Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6"]
+ Filter Operator [FIL_272] (rows=287989836 width=135)
+ predicate:(cs_catalog_page_sk is not null and cs_item_sk is not null and cs_promo_sk is not null and cs_sold_date_sk is not null)
+ TableScan [TS_39] (rows=287989836 width=135)
+ default@catalog_sales,catalog_sales,Tbl:COMPLETE,Col:NONE,Output:["cs_sold_date_sk","cs_catalog_page_sk","cs_item_sk","cs_promo_sk","cs_order_number","cs_ext_sales_price","cs_net_profit"]
+ <-Map 28 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_277]
+ PartitionCols:_col0, _col1
+ Select Operator [SEL_276] (rows=28798881 width=106)
+ Output:["_col0","_col1","_col2","_col3"]
+ Filter Operator [FIL_275] (rows=28798881 width=106)
+ predicate:cr_item_sk is not null
+ TableScan [TS_42] (rows=28798881 width=106)
+ default@catalog_returns,catalog_returns,Tbl:COMPLETE,Col:NONE,Output:["cr_item_sk","cr_order_number","cr_return_amount","cr_net_loss"]
<-Reducer 22 [CONTAINS] vectorized
- Reduce Output Operator [RS_292]
+ Reduce Output Operator [RS_299]
PartitionCols:_col0, _col1, _col2
- Group By Operator [GBY_291] (rows=2435062716 width=108)
+ Group By Operator [GBY_298] (rows=2435062716 width=108)
Output:["_col0","_col1","_col2","_col3","_col4","_col5"],aggregations:["sum(_col2)","sum(_col3)","sum(_col4)"],keys:_col0, _col1, 0L
- Select Operator [SEL_290] (rows=115958879 width=135)
- Output:["_col0","_col1","_col2","_col3","_col4"]
- Group By Operator [GBY_289] (rows=115958879 width=135)
- Output:["_col0","_col1","_col2","_col3"],aggregations:["sum(VALUE._col0)","sum(VALUE._col1)","sum(VALUE._col2)"],keys:KEY._col0
- <-Reducer 21 [SIMPLE_EDGE]
- SHUFFLE [RS_115]
- PartitionCols:_col0
- Group By Operator [GBY_114] (rows=231917759 width=135)
- Output:["_col0","_col1","_col2","_col3"],aggregations:["sum(_col1)","sum(_col2)","sum(_col3)"],keys:_col0
- Select Operator [SEL_112] (rows=231917759 width=135)
- Output:["_col0","_col1","_col2","_col3"]
- Merge Join Operator [MERGEJOIN_217] (rows=231917759 width=135)
- Conds:RS_109._col2=RS_288._col0(Inner),Output:["_col5","_col6","_col9","_col10","_col18"]
- <-Map 33 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_288]
- PartitionCols:_col0
- Select Operator [SEL_287] (rows=84 width=1850)
- Output:["_col0","_col1"]
- Filter Operator [FIL_286] (rows=84 width=1850)
- predicate:web_site_sk is not null
- TableScan [TS_94] (rows=84 width=1850)
- default@web_site,web_site,Tbl:COMPLETE,Col:NONE,Output:["web_site_sk","web_site_id"]
- <-Reducer 20 [SIMPLE_EDGE]
- SHUFFLE [RS_109]
- PartitionCols:_col2
- Merge Join Operator [MERGEJOIN_216] (rows=210834322 width=135)
- Conds:RS_106._col3=RS_253._col0(Inner),Output:["_col2","_col5","_col6","_col9","_col10"]
- <-Map 24 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_253]
- PartitionCols:_col0
- Please refer to the previous Select Operator [SEL_250]
- <-Reducer 19 [SIMPLE_EDGE]
- SHUFFLE [RS_106]
- PartitionCols:_col3
- Merge Join Operator [MERGEJOIN_215] (rows=191667562 width=135)
- Conds:RS_103._col1=RS_248._col0(Inner),Output:["_col2","_col3","_col5","_col6","_col9","_col10"]
- <-Map 23 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_248]
- PartitionCols:_col0
- Please refer to the previous Select Operator [SEL_245]
- <-Reducer 18 [SIMPLE_EDGE]
- SHUFFLE [RS_103]
- PartitionCols:_col1
- Merge Join Operator [MERGEJOIN_214] (rows=174243235 width=135)
- Conds:RS_100._col0=RS_243._col0(Inner),Output:["_col1","_col2","_col3","_col5","_col6","_col9","_col10"]
- <-Map 12 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_243]
- PartitionCols:_col0
- Please refer to the previous Select Operator [SEL_240]
- <-Reducer 31 [SIMPLE_EDGE]
- SHUFFLE [RS_100]
- PartitionCols:_col0
- Merge Join Operator [MERGEJOIN_213] (rows=158402938 width=135)
- Conds:RS_282._col1, _col4=RS_285._col0, _col1(Left Outer),Output:["_col0","_col1","_col2","_col3","_col5","_col6","_col9","_col10"]
- <-Map 30 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_282]
- PartitionCols:_col1, _col4
- Select Operator [SEL_281] (rows=144002668 width=135)
- Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6"]
- Filter Operator [FIL_280] (rows=144002668 width=135)
- predicate:(ws_item_sk is not null and ws_promo_sk is not null and ws_sold_date_sk is not null and ws_web_site_sk is not null)
- TableScan [TS_79] (rows=144002668 width=135)
- default@web_sales,web_sales,Tbl:COMPLETE,Col:NONE,Output:["ws_sold_date_sk","ws_item_sk","ws_web_site_sk","ws_promo_sk","ws_order_number","ws_ext_sales_price","ws_net_profit"]
- <-Map 32 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_285]
- PartitionCols:_col0, _col1
- Select Operator [SEL_284] (rows=14398467 width=92)
- Output:["_col0","_col1","_col2","_col3"]
- Filter Operator [FIL_283] (rows=14398467 width=92)
- predicate:wr_item_sk is not null
- TableScan [TS_82] (rows=14398467 width=92)
- default@web_returns,web_returns,Tbl:COMPLETE,Col:NONE,Output:["wr_item_sk","wr_order_number","wr_return_amt","wr_net_loss"]
+ Top N Key Operator [TNK_297] (rows=811687572 width=108)
+ keys:_col0, _col1, 0L,sort order:+++,top n:100
+ Select Operator [SEL_296] (rows=115958879 width=135)
+ Output:["_col0","_col1","_col2","_col3","_col4"]
+ Group By Operator [GBY_295] (rows=115958879 width=135)
+ Output:["_col0","_col1","_col2","_col3"],aggregations:["sum(VALUE._col0)","sum(VALUE._col1)","sum(VALUE._col2)"],keys:KEY._col0
+ <-Reducer 21 [SIMPLE_EDGE]
+ SHUFFLE [RS_115]
+ PartitionCols:_col0
+ Group By Operator [GBY_114] (rows=231917759 width=135)
+ Output:["_col0","_col1","_col2","_col3"],aggregations:["sum(_col1)","sum(_col2)","sum(_col3)"],keys:_col0
+ Select Operator [SEL_112] (rows=231917759 width=135)
+ Output:["_col0","_col1","_col2","_col3"]
+ Merge Join Operator [MERGEJOIN_218] (rows=231917759 width=135)
+ Conds:RS_109._col2=RS_294._col0(Inner),Output:["_col5","_col6","_col9","_col10","_col18"]
+ <-Map 33 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_294]
+ PartitionCols:_col0
+ Select Operator [SEL_293] (rows=84 width=1850)
+ Output:["_col0","_col1"]
+ Filter Operator [FIL_292] (rows=84 width=1850)
+ predicate:web_site_sk is not null
+ TableScan [TS_94] (rows=84 width=1850)
+ default@web_site,web_site,Tbl:COMPLETE,Col:NONE,Output:["web_site_sk","web_site_id"]
+ <-Reducer 20 [SIMPLE_EDGE]
+ SHUFFLE [RS_109]
+ PartitionCols:_col2
+ Merge Join Operator [MERGEJOIN_217] (rows=210834322 width=135)
+ Conds:RS_106._col3=RS_257._col0(Inner),Output:["_col2","_col5","_col6","_col9","_col10"]
+ <-Map 24 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_257]
+ PartitionCols:_col0
+ Please refer to the previous Select Operator [SEL_254]
+ <-Reducer 19 [SIMPLE_EDGE]
+ SHUFFLE [RS_106]
+ PartitionCols:_col3
+ Merge Join Operator [MERGEJOIN_216] (rows=191667562 width=135)
+ Conds:RS_103._col1=RS_252._col0(Inner),Output:["_col2","_col3","_col5","_col6","_col9","_col10"]
+ <-Map 23 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_252]
+ PartitionCols:_col0
+ Please refer to the previous Select Operator [SEL_249]
+ <-Reducer 18 [SIMPLE_EDGE]
+ SHUFFLE [RS_103]
+ PartitionCols:_col1
+ Merge Join Operator [MERGEJOIN_215] (rows=174243235 width=135)
+ Conds:RS_100._col0=RS_247._col0(Inner),Output:["_col1","_col2","_col3","_col5","_col6","_col9","_col10"]
+ <-Map 12 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_247]
+ PartitionCols:_col0
+ Please refer to the previous Select Operator [SEL_244]
+ <-Reducer 31 [SIMPLE_EDGE]
+ SHUFFLE [RS_100]
+ PartitionCols:_col0
+ Merge Join Operator [MERGEJOIN_214] (rows=158402938 width=135)
+ Conds:RS_288._col1, _col4=RS_291._col0, _col1(Left Outer),Output:["_col0","_col1","_col2","_col3","_col5","_col6","_col9","_col10"]
+ <-Map 30 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_288]
+ PartitionCols:_col1, _col4
+ Select Operator [SEL_287] (rows=144002668 width=135)
+ Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6"]
+ Filter Operator [FIL_286] (rows=144002668 width=135)
+ predicate:(ws_item_sk is not null and ws_promo_sk is not null and ws_sold_date_sk is not null and ws_web_site_sk is not null)
+ TableScan [TS_79] (rows=144002668 width=135)
+ default@web_sales,web_sales,Tbl:COMPLETE,Col:NONE,Output:["ws_sold_date_sk","ws_item_sk","ws_web_site_sk","ws_promo_sk","ws_order_number","ws_ext_sales_price","ws_net_profit"]
+ <-Map 32 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_291]
+ PartitionCols:_col0, _col1
+ Select Operator [SEL_290] (rows=14398467 width=92)
+ Output:["_col0","_col1","_col2","_col3"]
+ Filter Operator [FIL_289] (rows=14398467 width=92)
+ predicate:wr_item_sk is not null
+ TableScan [TS_82] (rows=14398467 width=92)
+ default@web_returns,web_returns,Tbl:COMPLETE,Col:NONE,Output:["wr_item_sk","wr_order_number","wr_return_amt","wr_net_loss"]
<-Reducer 7 [CONTAINS] vectorized
- Reduce Output Operator [RS_260]
+ Reduce Output Operator [RS_265]
PartitionCols:_col0, _col1, _col2
- Group By Operator [GBY_259] (rows=2435062716 width=108)
+ Group By Operator [GBY_264] (rows=2435062716 width=108)
Output:["_col0","_col1","_col2","_col3","_col4","_col5"],aggregations:["sum(_col2)","sum(_col3)","sum(_col4)"],keys:_col0, _col1, 0L
- Select Operator [SEL_258] (rows=463823414 width=88)
- Output:["_col0","_col1","_col2","_col3","_col4"]
- Group By Operator [GBY_257] (rows=463823414 width=88)
- Output:["_col0","_col1","_col2","_col3"],aggregations:["sum(VALUE._col0)","sum(VALUE._col1)","sum(VALUE._col2)"],keys:KEY._col0
- <-Reducer 6 [SIMPLE_EDGE]
- SHUFFLE [RS_36]
- PartitionCols:_col0
- Group By Operator [GBY_35] (rows=927646829 width=88)
- Output:["_col0","_col1","_col2","_col3"],aggregations:["sum(_col1)","sum(_col2)","sum(_col3)"],keys:_col0
- Select Operator [SEL_33] (rows=927646829 width=88)
- Output:["_col0","_col1","_col2","_col3"]
- Merge Join Operator [MERGEJOIN_207] (rows=927646829 width=88)
- Conds:RS_30._col2=RS_256._col0(Inner),Output:["_col5","_col6","_col9","_col10","_col18"]
- <-Map 25 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_256]
- PartitionCols:_col0
- Select Operator [SEL_255] (rows=1704 width=1910)
- Output:["_col0","_col1"]
- Filter Operator [FIL_254] (rows=1704 width=1910)
- predicate:s_store_sk is not null
- TableScan [TS_15] (rows=1704 width=1910)
- default@store,store,Tbl:COMPLETE,Col:NONE,Output:["s_store_sk","s_store_id"]
- <-Reducer 5 [SIMPLE_EDGE]
- SHUFFLE [RS_30]
- PartitionCols:_col2
- Merge Join Operator [MERGEJOIN_206] (rows=843315281 width=88)
- Conds:RS_27._col3=RS_251._col0(Inner),Output:["_col2","_col5","_col6","_col9","_col10"]
- <-Map 24 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_251]
- PartitionCols:_col0
- Please refer to the previous Select Operator [SEL_250]
- <-Reducer 4 [SIMPLE_EDGE]
- SHUFFLE [RS_27]
- PartitionCols:_col3
- Merge Join Operator [MERGEJOIN_205] (rows=766650239 width=88)
- Conds:RS_24._col1=RS_246._col0(Inner),Output:["_col2","_col3","_col5","_col6","_col9","_col10"]
- <-Map 23 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_246]
- PartitionCols:_col0
- Please refer to the previous Select Operator [SEL_245]
- <-Reducer 3 [SIMPLE_EDGE]
- SHUFFLE [RS_24]
- PartitionCols:_col1
- Merge Join Operator [MERGEJOIN_204] (rows=696954748 width=88)
- Conds:RS_21._col0=RS_241._col0(Inner),Output:["_col1","_col2","_col3","_col5","_col6","_col9","_col10"]
- <-Map 12 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_241]
- PartitionCols:_col0
- Please refer to the previous Select Operator [SEL_240]
- <-Reducer 2 [SIMPLE_EDGE]
- SHUFFLE [RS_21]
- PartitionCols:_col0
- Merge Join Operator [MERGEJOIN_203] (rows=633595212 width=88)
- Conds:RS_235._col1, _col4=RS_238._col0, _col1(Left Outer),Output:["_col0","_col1","_col2","_col3","_col5","_col6","_col9","_col10"]
- <-Map 1 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_235]
- PartitionCols:_col1, _col4
- Select Operator [SEL_234] (rows=575995635 width=88)
- Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6"]
- Filter Operator [FIL_233] (rows=575995635 width=88)
- predicate:(ss_item_sk is not null and ss_promo_sk is not null and ss_sold_date_sk is not null and ss_store_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_store_sk","ss_promo_sk","ss_ticket_number","ss_ext_sales_price","ss_net_profit"]
- <-Map 11 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_238]
- PartitionCols:_col0, _col1
- Select Operator [SEL_237] (rows=57591150 width=77)
- Output:["_col0","_col1","_col2","_col3"]
- Filter Operator [FIL_236] (rows=57591150 width=77)
- predicate:sr_item_sk is not null
- TableScan [TS_3] (rows=57591150 width=77)
- default@store_returns,store_returns,Tbl:COMPLETE,Col:NONE,Output:["sr_item_sk","sr_ticket_number","sr_return_amt","sr_net_loss"]
+ Top N Key Operator [TNK_263] (rows=811687572 width=108)
+ keys:_col0, _col1, 0L,sort order:+++,top n:100
+ Select Operator [SEL_262] (rows=463823414 width=88)
+ Output:["_col0","_col1","_col2","_col3","_col4"]
+ Group By Operator [GBY_261] (rows=463823414 width=88)
+ Output:["_col0","_col1","_col2","_col3"],aggregations:["sum(VALUE._col0)","sum(VALUE._col1)","sum(VALUE._col2)"],keys:KEY._col0
+ <-Reducer 6 [SIMPLE_EDGE]
+ SHUFFLE [RS_36]
+ PartitionCols:_col0
+ Group By Operator [GBY_35] (rows=927646829 width=88)
+ Output:["_col0","_col1","_col2","_col3"],aggregations:["sum(_col1)","sum(_col2)","sum(_col3)"],keys:_col0
+ Select Operator [SEL_33] (rows=927646829 width=88)
+ Output:["_col0","_col1","_col2","_col3"]
+ Merge Join Operator [MERGEJOIN_208] (rows=927646829 width=88)
+ Conds:RS_30._col2=RS_260._col0(Inner),Output:["_col5","_col6","_col9","_col10","_col18"]
+ <-Map 25 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_260]
+ PartitionCols:_col0
+ Select Operator [SEL_259] (rows=1704 width=1910)
+ Output:["_col0","_col1"]
+ Filter Operator [FIL_258] (rows=1704 width=1910)
+ predicate:s_store_sk is not null
+ TableScan [TS_15] (rows=1704 width=1910)
+ default@store,store,Tbl:COMPLETE,Col:NONE,Output:["s_store_sk","s_store_id"]
+ <-Reducer 5 [SIMPLE_EDGE]
+ SHUFFLE [RS_30]
+ PartitionCols:_col2
+ Merge Join Operator [MERGEJOIN_207] (rows=843315281 width=88)
+ Conds:RS_27._col3=RS_255._col0(Inner),Output:["_col2","_col5","_col6","_col9","_col10"]
+ <-Map 24 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_255]
+ PartitionCols:_col0
+ Please refer to the previous Select Operator [SEL_254]
+ <-Reducer 4 [SIMPLE_EDGE]
+ SHUFFLE [RS_27]
+ PartitionCols:_col3
+ Merge Join Operator [MERGEJOIN_206] (rows=766650239 width=88)
+ Conds:RS_24._col1=RS_250._col0(Inner),Output:["_col2","_col3","_col5","_col6","_col9","_col10"]
+ <-Map 23 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_250]
+ PartitionCols:_col0
+ Please refer to the previous Select Operator [SEL_249]
+ <-Reducer 3 [SIMPLE_EDGE]
+ SHUFFLE [RS_24]
+ PartitionCols:_col1
+ Merge Join Operator [MERGEJOIN_205] (rows=696954748 width=88)
+ Conds:RS_21._col0=RS_245._col0(Inner),Output:["_col1","_col2","_col3","_col5","_col6","_col9","_col10"]
+ <-Map 12 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_245]
+ PartitionCols:_col0
+ Please refer to the previous Select Operator [SEL_244]
+ <-Reducer 2 [SIMPLE_EDGE]
+ SHUFFLE [RS_21]
+ PartitionCols:_col0
+ Merge Join Operator [MERGEJOIN_204] (rows=633595212 width=88)
+ Conds:RS_239._col1, _col4=RS_242._col0, _col1(Left Outer),Output:["_col0","_col1","_col2","_col3","_col5","_col6","_col9","_col10"]
+ <-Map 1 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_239]
+ PartitionCols:_col1, _col4
+ Select Operator [SEL_238] (rows=575995635 width=88)
+ Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6"]
+ Filter Operator [FIL_237] (rows=575995635 width=88)
+ predicate:(ss_item_sk is not null and ss_promo_sk is not null and ss_sold_date_sk is not null and ss_store_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_store_sk","ss_promo_sk","ss_ticket_number","ss_ext_sales_price","ss_net_profit"]
+ <-Map 11 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_242]
+ PartitionCols:_col0, _col1
+ Select Operator [SEL_241] (rows=57591150 width=77)
+ Output:["_col0","_col1","_col2","_col3"]
+ Filter Operator [FIL_240] (rows=57591150 width=77)
+ predicate:sr_item_sk is not null
+ TableScan [TS_3] (rows=57591150 width=77)
+ default@store_returns,store_returns,Tbl:COMPLETE,Col:NONE,Output:["sr_item_sk","sr_ticket_number","sr_return_amt","sr_net_loss"]
http://git-wip-us.apache.org/repos/asf/hive/blob/cc294d32/ql/src/test/results/clientpositive/perf/tez/query82.q.out
----------------------------------------------------------------------
diff --git a/ql/src/test/results/clientpositive/perf/tez/query82.q.out b/ql/src/test/results/clientpositive/perf/tez/query82.q.out
index 9a9b707..d948b7d 100644
--- a/ql/src/test/results/clientpositive/perf/tez/query82.q.out
+++ b/ql/src/test/results/clientpositive/perf/tez/query82.q.out
@@ -43,61 +43,63 @@ Stage-0
limit:100
Stage-1
Reducer 4 vectorized
- File Output Operator [FS_61]
- Limit [LIM_60] (rows=100 width=88)
+ File Output Operator [FS_62]
+ Limit [LIM_61] (rows=100 width=88)
Number of rows:100
- Select Operator [SEL_59] (rows=633595212 width=88)
+ Select Operator [SEL_60] (rows=633595212 width=88)
Output:["_col0","_col1","_col2"]
<-Reducer 3 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_58]
- Group By Operator [GBY_57] (rows=633595212 width=88)
+ SHUFFLE [RS_59]
+ Group By Operator [GBY_58] (rows=633595212 width=88)
Output:["_col0","_col1","_col2"],keys:KEY._col0, KEY._col1, KEY._col2
<-Reducer 2 [SIMPLE_EDGE]
SHUFFLE [RS_22]
PartitionCols:_col0, _col1, _col2
Group By Operator [GBY_21] (rows=1267190424 width=88)
Output:["_col0","_col1","_col2"],keys:_col2, _col3, _col4
- Merge Join Operator [MERGEJOIN_44] (rows=1267190424 width=88)
- Conds:RS_47._col0=RS_50._col0(Inner),RS_50._col0=RS_18._col1(Inner),Output:["_col2","_col3","_col4"]
- <-Map 1 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_47]
- PartitionCols:_col0
- Select Operator [SEL_46] (rows=575995635 width=88)
- Output:["_col0"]
- Filter Operator [FIL_45] (rows=575995635 width=88)
- predicate:ss_item_sk is not null
- TableScan [TS_0] (rows=575995635 width=88)
- default@store_sales,store_sales,Tbl:COMPLETE,Col:NONE,Output:["ss_item_sk"]
- <-Map 5 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_50]
- PartitionCols:_col0
- Select Operator [SEL_49] (rows=25666 width=1436)
- Output:["_col0","_col1","_col2","_col3"]
- Filter Operator [FIL_48] (rows=25666 width=1436)
- predicate:((i_manufact_id) IN (437, 129, 727, 663) and i_current_price BETWEEN 30 AND 60 and i_item_sk is not null)
- TableScan [TS_3] (rows=462000 width=1436)
- default@item,item,Tbl:COMPLETE,Col:NONE,Output:["i_item_sk","i_item_id","i_item_desc","i_current_price","i_manufact_id"]
- <-Reducer 7 [SIMPLE_EDGE]
- SHUFFLE [RS_18]
- PartitionCols:_col1
- Merge Join Operator [MERGEJOIN_43] (rows=4593600 width=15)
- Conds:RS_53._col0=RS_56._col0(Inner),Output:["_col1"]
- <-Map 6 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_53]
- PartitionCols:_col0
- Select Operator [SEL_52] (rows=4176000 width=15)
- Output:["_col0","_col1"]
- Filter Operator [FIL_51] (rows=4176000 width=15)
- predicate:(inv_date_sk is not null and inv_item_sk is not null and inv_quantity_on_hand BETWEEN 100 AND 500)
- TableScan [TS_6] (rows=37584000 width=15)
- default@inventory,inventory,Tbl:COMPLETE,Col:NONE,Output:["inv_date_sk","inv_item_sk","inv_quantity_on_hand"]
- <-Map 8 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_56]
- PartitionCols:_col0
- Select Operator [SEL_55] (rows=8116 width=1119)
- Output:["_col0"]
- Filter Operator [FIL_54] (rows=8116 width=1119)
- predicate:(CAST( d_date AS TIMESTAMP) BETWEEN TIMESTAMP'2002-05-30 00:00:00' AND TIMESTAMP'2002-07-29 00:00:00' and d_date_sk is not null)
- TableScan [TS_9] (rows=73049 width=1119)
- default@date_dim,date_dim,Tbl:COMPLETE,Col:NONE,Output:["d_date_sk","d_date"]
+ Top N Key Operator [TNK_43] (rows=1267190424 width=88)
+ keys:_col2, _col3, _col4,sort order:+++,top n:100
+ Merge Join Operator [MERGEJOIN_45] (rows=1267190424 width=88)
+ Conds:RS_48._col0=RS_51._col0(Inner),RS_51._col0=RS_18._col1(Inner),Output:["_col2","_col3","_col4"]
+ <-Map 1 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_48]
+ PartitionCols:_col0
+ Select Operator [SEL_47] (rows=575995635 width=88)
+ Output:["_col0"]
+ Filter Operator [FIL_46] (rows=575995635 width=88)
+ predicate:ss_item_sk is not null
+ TableScan [TS_0] (rows=575995635 width=88)
+ default@store_sales,store_sales,Tbl:COMPLETE,Col:NONE,Output:["ss_item_sk"]
+ <-Map 5 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_51]
+ PartitionCols:_col0
+ Select Operator [SEL_50] (rows=25666 width=1436)
+ Output:["_col0","_col1","_col2","_col3"]
+ Filter Operator [FIL_49] (rows=25666 width=1436)
+ predicate:((i_manufact_id) IN (437, 129, 727, 663) and i_current_price BETWEEN 30 AND 60 and i_item_sk is not null)
+ TableScan [TS_3] (rows=462000 width=1436)
+ default@item,item,Tbl:COMPLETE,Col:NONE,Output:["i_item_sk","i_item_id","i_item_desc","i_current_price","i_manufact_id"]
+ <-Reducer 7 [SIMPLE_EDGE]
+ SHUFFLE [RS_18]
+ PartitionCols:_col1
+ Merge Join Operator [MERGEJOIN_44] (rows=4593600 width=15)
+ Conds:RS_54._col0=RS_57._col0(Inner),Output:["_col1"]
+ <-Map 6 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_54]
+ PartitionCols:_col0
+ Select Operator [SEL_53] (rows=4176000 width=15)
+ Output:["_col0","_col1"]
+ Filter Operator [FIL_52] (rows=4176000 width=15)
+ predicate:(inv_date_sk is not null and inv_item_sk is not null and inv_quantity_on_hand BETWEEN 100 AND 500)
+ TableScan [TS_6] (rows=37584000 width=15)
+ default@inventory,inventory,Tbl:COMPLETE,Col:NONE,Output:["inv_date_sk","inv_item_sk","inv_quantity_on_hand"]
+ <-Map 8 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_57]
+ PartitionCols:_col0
+ Select Operator [SEL_56] (rows=8116 width=1119)
+ Output:["_col0"]
+ Filter Operator [FIL_55] (rows=8116 width=1119)
+ predicate:(CAST( d_date AS TIMESTAMP) BETWEEN TIMESTAMP'2002-05-30 00:00:00' AND TIMESTAMP'2002-07-29 00:00:00' and d_date_sk is not null)
+ TableScan [TS_9] (rows=73049 width=1119)
+ default@date_dim,date_dim,Tbl:COMPLETE,Col:NONE,Output:["d_date_sk","d_date"]
http://git-wip-us.apache.org/repos/asf/hive/blob/cc294d32/ql/src/test/results/clientpositive/perf/tez/query99.q.out
----------------------------------------------------------------------
diff --git a/ql/src/test/results/clientpositive/perf/tez/query99.q.out b/ql/src/test/results/clientpositive/perf/tez/query99.q.out
index 32fb96e..ec14e0a 100644
--- a/ql/src/test/results/clientpositive/perf/tez/query99.q.out
+++ b/ql/src/test/results/clientpositive/perf/tez/query99.q.out
@@ -81,84 +81,86 @@ Stage-0
limit:-1
Stage-1
Reducer 7 vectorized
- File Output Operator [FS_81]
- Limit [LIM_80] (rows=100 width=135)
+ File Output Operator [FS_82]
+ Limit [LIM_81] (rows=100 width=135)
Number of rows:100
- Select Operator [SEL_79] (rows=210822976 width=135)
+ Select Operator [SEL_80] (rows=210822976 width=135)
Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6","_col7"]
<-Reducer 6 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_78]
- Select Operator [SEL_77] (rows=210822976 width=135)
+ SHUFFLE [RS_79]
+ Select Operator [SEL_78] (rows=210822976 width=135)
Output:["_col1","_col2","_col3","_col4","_col5","_col6","_col7","_col8"]
- Group By Operator [GBY_76] (rows=210822976 width=135)
+ Group By Operator [GBY_77] (rows=210822976 width=135)
Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6","_col7"],aggregations:["sum(VALUE._col0)","sum(VALUE._col1)","sum(VALUE._col2)","sum(VALUE._col3)","sum(VALUE._col4)"],keys:KEY._col0, KEY._col1, KEY._col2
<-Reducer 5 [SIMPLE_EDGE]
SHUFFLE [RS_30]
PartitionCols:_col0, _col1, _col2
Group By Operator [GBY_29] (rows=421645953 width=135)
Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6","_col7"],aggregations:["sum(_col3)","sum(_col4)","sum(_col5)","sum(_col6)","sum(_col7)"],keys:_col0, _col1, _col2
- Select Operator [SEL_27] (rows=421645953 width=135)
- Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6","_col7"]
- Merge Join Operator [MERGEJOIN_60] (rows=421645953 width=135)
- Conds:RS_24._col3=RS_75._col0(Inner),Output:["_col0","_col1","_col8","_col10","_col12"]
- <-Map 11 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_75]
- PartitionCols:_col0
- Select Operator [SEL_74] (rows=1 width=0)
- Output:["_col0","_col1"]
- Filter Operator [FIL_73] (rows=1 width=0)
- predicate:sm_ship_mode_sk is not null
- TableScan [TS_12] (rows=1 width=0)
- default@ship_mode,ship_mode,Tbl:PARTIAL,Col:NONE,Output:["sm_ship_mode_sk","sm_type"]
- <-Reducer 4 [SIMPLE_EDGE]
- SHUFFLE [RS_24]
- PartitionCols:_col3
- Merge Join Operator [MERGEJOIN_59] (rows=383314495 width=135)
- Conds:RS_21._col4=RS_72._col0(Inner),Output:["_col0","_col1","_col3","_col8","_col10"]
- <-Map 10 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_72]
- PartitionCols:_col0
- Select Operator [SEL_71] (rows=27 width=1029)
- Output:["_col0","_col1"]
- Filter Operator [FIL_70] (rows=27 width=1029)
- predicate:w_warehouse_sk is not null
- TableScan [TS_9] (rows=27 width=1029)
- default@warehouse,warehouse,Tbl:COMPLETE,Col:NONE,Output:["w_warehouse_sk","w_warehouse_name"]
- <-Reducer 3 [SIMPLE_EDGE]
- SHUFFLE [RS_21]
- PartitionCols:_col4
- Merge Join Operator [MERGEJOIN_58] (rows=348467716 width=135)
- Conds:RS_18._col2=RS_69._col0(Inner),Output:["_col0","_col1","_col3","_col4","_col8"]
- <-Map 9 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_69]
- PartitionCols:_col0
- Select Operator [SEL_68] (rows=60 width=2045)
- Output:["_col0","_col1"]
- Filter Operator [FIL_67] (rows=60 width=2045)
- predicate:cc_call_center_sk is not null
- TableScan [TS_6] (rows=60 width=2045)
- default@call_center,call_center,Tbl:COMPLETE,Col:NONE,Output:["cc_call_center_sk","cc_name"]
- <-Reducer 2 [SIMPLE_EDGE]
- SHUFFLE [RS_18]
- PartitionCols:_col2
- Merge Join Operator [MERGEJOIN_57] (rows=316788826 width=135)
- Conds:RS_63._col1=RS_66._col0(Inner),Output:["_col0","_col1","_col2","_col3","_col4"]
- <-Map 1 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_63]
- PartitionCols:_col1
- Select Operator [SEL_62] (rows=287989836 width=135)
- Output:["_col0","_col1","_col2","_col3","_col4"]
- Filter Operator [FIL_61] (rows=287989836 width=135)
- predicate:(cs_call_center_sk is not null and cs_ship_date_sk is not null and cs_ship_mode_sk is not null and cs_warehouse_sk is not null)
- TableScan [TS_0] (rows=287989836 width=135)
- default@catalog_sales,catalog_sales,Tbl:COMPLETE,Col:NONE,Output:["cs_sold_date_sk","cs_ship_date_sk","cs_call_center_sk","cs_ship_mode_sk","cs_warehouse_sk"]
- <-Map 8 [SIMPLE_EDGE] vectorized
- SHUFFLE [RS_66]
- PartitionCols:_col0
- Select Operator [SEL_65] (rows=8116 width=1119)
- Output:["_col0"]
- Filter Operator [FIL_64] (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_month_seq"]
+ Top N Key Operator [TNK_57] (rows=421645953 width=135)
+ keys:_col0, _col1, _col2,sort order:+++,top n:100
+ Select Operator [SEL_27] (rows=421645953 width=135)
+ Output:["_col0","_col1","_col2","_col3","_col4","_col5","_col6","_col7"]
+ Merge Join Operator [MERGEJOIN_61] (rows=421645953 width=135)
+ Conds:RS_24._col3=RS_76._col0(Inner),Output:["_col0","_col1","_col8","_col10","_col12"]
+ <-Map 11 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_76]
+ PartitionCols:_col0
+ Select Operator [SEL_75] (rows=1 width=0)
+ Output:["_col0","_col1"]
+ Filter Operator [FIL_74] (rows=1 width=0)
+ predicate:sm_ship_mode_sk is not null
+ TableScan [TS_12] (rows=1 width=0)
+ default@ship_mode,ship_mode,Tbl:PARTIAL,Col:NONE,Output:["sm_ship_mode_sk","sm_type"]
+ <-Reducer 4 [SIMPLE_EDGE]
+ SHUFFLE [RS_24]
+ PartitionCols:_col3
+ Merge Join Operator [MERGEJOIN_60] (rows=383314495 width=135)
+ Conds:RS_21._col4=RS_73._col0(Inner),Output:["_col0","_col1","_col3","_col8","_col10"]
+ <-Map 10 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_73]
+ PartitionCols:_col0
+ Select Operator [SEL_72] (rows=27 width=1029)
+ Output:["_col0","_col1"]
+ Filter Operator [FIL_71] (rows=27 width=1029)
+ predicate:w_warehouse_sk is not null
+ TableScan [TS_9] (rows=27 width=1029)
+ default@warehouse,warehouse,Tbl:COMPLETE,Col:NONE,Output:["w_warehouse_sk","w_warehouse_name"]
+ <-Reducer 3 [SIMPLE_EDGE]
+ SHUFFLE [RS_21]
+ PartitionCols:_col4
+ Merge Join Operator [MERGEJOIN_59] (rows=348467716 width=135)
+ Conds:RS_18._col2=RS_70._col0(Inner),Output:["_col0","_col1","_col3","_col4","_col8"]
+ <-Map 9 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_70]
+ PartitionCols:_col0
+ Select Operator [SEL_69] (rows=60 width=2045)
+ Output:["_col0","_col1"]
+ Filter Operator [FIL_68] (rows=60 width=2045)
+ predicate:cc_call_center_sk is not null
+ TableScan [TS_6] (rows=60 width=2045)
+ default@call_center,call_center,Tbl:COMPLETE,Col:NONE,Output:["cc_call_center_sk","cc_name"]
+ <-Reducer 2 [SIMPLE_EDGE]
+ SHUFFLE [RS_18]
+ PartitionCols:_col2
+ Merge Join Operator [MERGEJOIN_58] (rows=316788826 width=135)
+ Conds:RS_64._col1=RS_67._col0(Inner),Output:["_col0","_col1","_col2","_col3","_col4"]
+ <-Map 1 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_64]
+ PartitionCols:_col1
+ Select Operator [SEL_63] (rows=287989836 width=135)
+ Output:["_col0","_col1","_col2","_col3","_col4"]
+ Filter Operator [FIL_62] (rows=287989836 width=135)
+ predicate:(cs_call_center_sk is not null and cs_ship_date_sk is not null and cs_ship_mode_sk is not null and cs_warehouse_sk is not null)
+ TableScan [TS_0] (rows=287989836 width=135)
+ default@catalog_sales,catalog_sales,Tbl:COMPLETE,Col:NONE,Output:["cs_sold_date_sk","cs_ship_date_sk","cs_call_center_sk","cs_ship_mode_sk","cs_warehouse_sk"]
+ <-Map 8 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_67]
+ PartitionCols:_col0
+ Select Operator [SEL_66] (rows=8116 width=1119)
+ Output:["_col0"]
+ Filter Operator [FIL_65] (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_month_seq"]
http://git-wip-us.apache.org/repos/asf/hive/blob/cc294d32/ql/src/test/results/clientpositive/tez/topnkey.q.out
----------------------------------------------------------------------
diff --git a/ql/src/test/results/clientpositive/tez/topnkey.q.out b/ql/src/test/results/clientpositive/tez/topnkey.q.out
new file mode 100644
index 0000000..66b9191
--- /dev/null
+++ b/ql/src/test/results/clientpositive/tez/topnkey.q.out
@@ -0,0 +1,162 @@
+PREHOOK: query: EXPLAIN
+SELECT key, SUM(CAST(SUBSTR(value,5) AS INT)) FROM src GROUP BY key ORDER BY key LIMIT 5
+PREHOOK: type: QUERY
+POSTHOOK: query: EXPLAIN
+SELECT key, SUM(CAST(SUBSTR(value,5) AS INT)) FROM src GROUP BY key ORDER BY key LIMIT 5
+POSTHOOK: type: QUERY
+Plan optimized by CBO.
+
+Vertex dependency in root stage
+Reducer 2 <- Map 1 (SIMPLE_EDGE)
+Reducer 3 <- Reducer 2 (SIMPLE_EDGE)
+
+Stage-0
+ Fetch Operator
+ limit:5
+ Stage-1
+ Reducer 3
+ File Output Operator [FS_10]
+ Limit [LIM_9] (rows=5 width=95)
+ Number of rows:5
+ Select Operator [SEL_8] (rows=250 width=95)
+ Output:["_col0","_col1"]
+ <-Reducer 2 [SIMPLE_EDGE]
+ SHUFFLE [RS_7]
+ Group By Operator [GBY_5] (rows=250 width=95)
+ Output:["_col0","_col1"],aggregations:["sum(VALUE._col0)"],keys:KEY._col0
+ <-Map 1 [SIMPLE_EDGE]
+ SHUFFLE [RS_4]
+ PartitionCols:_col0
+ Group By Operator [GBY_3] (rows=250 width=95)
+ Output:["_col0","_col1"],aggregations:["sum(_col1)"],keys:_col0
+ Top N Key Operator [TNK_11] (rows=500 width=178)
+ keys:_col0,sort order:+,top n:5
+ Select Operator [SEL_1] (rows=500 width=178)
+ Output:["_col0","_col1"]
+ TableScan [TS_0] (rows=500 width=178)
+ default@src,src,Tbl:COMPLETE,Col:COMPLETE,Output:["key","value"]
+
+PREHOOK: query: SELECT key, SUM(CAST(SUBSTR(value,5) AS INT)) FROM src GROUP BY key ORDER BY key LIMIT 5
+PREHOOK: type: QUERY
+PREHOOK: Input: default@src
+PREHOOK: Output: hdfs://### HDFS PATH ###
+POSTHOOK: query: SELECT key, SUM(CAST(SUBSTR(value,5) AS INT)) FROM src GROUP BY key ORDER BY key LIMIT 5
+POSTHOOK: type: QUERY
+POSTHOOK: Input: default@src
+POSTHOOK: Output: hdfs://### HDFS PATH ###
+0 0
+10 10
+100 200
+103 206
+104 208
+PREHOOK: query: EXPLAIN
+SELECT key FROM src GROUP BY key ORDER BY key LIMIT 5
+PREHOOK: type: QUERY
+POSTHOOK: query: EXPLAIN
+SELECT key FROM src GROUP BY key ORDER BY key LIMIT 5
+POSTHOOK: type: QUERY
+Plan optimized by CBO.
+
+Vertex dependency in root stage
+Reducer 2 <- Map 1 (SIMPLE_EDGE)
+Reducer 3 <- Reducer 2 (SIMPLE_EDGE)
+
+Stage-0
+ Fetch Operator
+ limit:5
+ Stage-1
+ Reducer 3
+ File Output Operator [FS_9]
+ Limit [LIM_8] (rows=5 width=87)
+ Number of rows:5
+ Select Operator [SEL_7] (rows=250 width=87)
+ Output:["_col0"]
+ <-Reducer 2 [SIMPLE_EDGE]
+ SHUFFLE [RS_6]
+ Group By Operator [GBY_4] (rows=250 width=87)
+ Output:["_col0"],keys:KEY._col0
+ <-Map 1 [SIMPLE_EDGE]
+ SHUFFLE [RS_3]
+ PartitionCols:_col0
+ Group By Operator [GBY_2] (rows=250 width=87)
+ Output:["_col0"],keys:key
+ Top N Key Operator [TNK_10] (rows=500 width=87)
+ keys:key,sort order:+,top n:5
+ Select Operator [SEL_1] (rows=500 width=87)
+ Output:["key"]
+ TableScan [TS_0] (rows=500 width=87)
+ default@src,src,Tbl:COMPLETE,Col:COMPLETE,Output:["key"]
+
+PREHOOK: query: SELECT key FROM src GROUP BY key ORDER BY key LIMIT 5
+PREHOOK: type: QUERY
+PREHOOK: Input: default@src
+PREHOOK: Output: hdfs://### HDFS PATH ###
+POSTHOOK: query: SELECT key FROM src GROUP BY key ORDER BY key LIMIT 5
+POSTHOOK: type: QUERY
+POSTHOOK: Input: default@src
+POSTHOOK: Output: hdfs://### HDFS PATH ###
+0
+10
+100
+103
+104
+PREHOOK: query: explain vectorization detail
+SELECT src1.key, src2.value FROM src src1 JOIN src src2 ON (src1.key = src2.key) ORDER BY src1.key LIMIT 5
+PREHOOK: type: QUERY
+POSTHOOK: query: explain vectorization detail
+SELECT src1.key, src2.value FROM src src1 JOIN src src2 ON (src1.key = src2.key) ORDER BY src1.key LIMIT 5
+POSTHOOK: type: QUERY
+Plan optimized by CBO.
+
+Vertex dependency in root stage
+Reducer 2 <- Map 1 (SIMPLE_EDGE), Map 4 (SIMPLE_EDGE)
+Reducer 3 <- Reducer 2 (SIMPLE_EDGE)
+
+Stage-0
+ Fetch Operator
+ limit:5
+ Stage-1
+ Reducer 3
+ File Output Operator [FS_13]
+ Limit [LIM_12] (rows=5 width=178)
+ Number of rows:5
+ Select Operator [SEL_11] (rows=791 width=178)
+ Output:["_col0","_col1"]
+ <-Reducer 2 [SIMPLE_EDGE]
+ SHUFFLE [RS_10]
+ Select Operator [SEL_9] (rows=791 width=178)
+ Output:["_col0","_col1"]
+ Merge Join Operator [MERGEJOIN_28] (rows=791 width=178)
+ Conds:RS_6._col0=RS_7._col0(Inner),Output:["_col0","_col2"]
+ <-Map 1 [SIMPLE_EDGE]
+ SHUFFLE [RS_6]
+ PartitionCols:_col0
+ Select Operator [SEL_2] (rows=500 width=87)
+ Output:["_col0"]
+ Filter Operator [FIL_16] (rows=500 width=87)
+ predicate:key is not null
+ TableScan [TS_0] (rows=500 width=87)
+ default@src,src1,Tbl:COMPLETE,Col:COMPLETE,Output:["key"]
+ <-Map 4 [SIMPLE_EDGE]
+ SHUFFLE [RS_7]
+ PartitionCols:_col0
+ Select Operator [SEL_5] (rows=500 width=178)
+ Output:["_col0","_col1"]
+ Filter Operator [FIL_17] (rows=500 width=178)
+ predicate:key is not null
+ TableScan [TS_3] (rows=500 width=178)
+ default@src,src2,Tbl:COMPLETE,Col:COMPLETE,Output:["key","value"]
+
+PREHOOK: query: SELECT src1.key, src2.value FROM src src1 JOIN src src2 ON (src1.key = src2.key) ORDER BY src1.key LIMIT 5
+PREHOOK: type: QUERY
+PREHOOK: Input: default@src
+PREHOOK: Output: hdfs://### HDFS PATH ###
+POSTHOOK: query: SELECT src1.key, src2.value FROM src src1 JOIN src src2 ON (src1.key = src2.key) ORDER BY src1.key LIMIT 5
+POSTHOOK: type: QUERY
+POSTHOOK: Input: default@src
+POSTHOOK: Output: hdfs://### HDFS PATH ###
+0 val_0
+0 val_0
+0 val_0
+0 val_0
+0 val_0
http://git-wip-us.apache.org/repos/asf/hive/blob/cc294d32/ql/src/test/results/clientpositive/tez/vector_topnkey.q.out
----------------------------------------------------------------------
diff --git a/ql/src/test/results/clientpositive/tez/vector_topnkey.q.out b/ql/src/test/results/clientpositive/tez/vector_topnkey.q.out
new file mode 100644
index 0000000..d6f7cc2
--- /dev/null
+++ b/ql/src/test/results/clientpositive/tez/vector_topnkey.q.out
@@ -0,0 +1,162 @@
+PREHOOK: query: explain vectorization detail
+SELECT key, SUM(CAST(SUBSTR(value,5) AS INT)) FROM src GROUP BY key ORDER BY key LIMIT 5
+PREHOOK: type: QUERY
+POSTHOOK: query: explain vectorization detail
+SELECT key, SUM(CAST(SUBSTR(value,5) AS INT)) FROM src GROUP BY key ORDER BY key LIMIT 5
+POSTHOOK: type: QUERY
+Plan optimized by CBO.
+
+Vertex dependency in root stage
+Reducer 2 <- Map 1 (SIMPLE_EDGE)
+Reducer 3 <- Reducer 2 (SIMPLE_EDGE)
+
+Stage-0
+ Fetch Operator
+ limit:5
+ Stage-1
+ Reducer 3 vectorized
+ File Output Operator [FS_20]
+ Limit [LIM_19] (rows=5 width=95)
+ Number of rows:5
+ Select Operator [SEL_18] (rows=250 width=95)
+ Output:["_col0","_col1"]
+ <-Reducer 2 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_17]
+ Group By Operator [GBY_16] (rows=250 width=95)
+ Output:["_col0","_col1"],aggregations:["sum(VALUE._col0)"],keys:KEY._col0
+ <-Map 1 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_15]
+ PartitionCols:_col0
+ Group By Operator [GBY_14] (rows=250 width=95)
+ Output:["_col0","_col1"],aggregations:["sum(_col1)"],keys:_col0
+ Top N Key Operator [TNK_13] (rows=500 width=178)
+ keys:_col0,sort order:+,top n:5
+ Select Operator [SEL_12] (rows=500 width=178)
+ Output:["_col0","_col1"]
+ TableScan [TS_0] (rows=500 width=178)
+ default@src,src,Tbl:COMPLETE,Col:COMPLETE,Output:["key","value"]
+
+PREHOOK: query: SELECT key, SUM(CAST(SUBSTR(value,5) AS INT)) FROM src GROUP BY key ORDER BY key LIMIT 5
+PREHOOK: type: QUERY
+PREHOOK: Input: default@src
+PREHOOK: Output: hdfs://### HDFS PATH ###
+POSTHOOK: query: SELECT key, SUM(CAST(SUBSTR(value,5) AS INT)) FROM src GROUP BY key ORDER BY key LIMIT 5
+POSTHOOK: type: QUERY
+POSTHOOK: Input: default@src
+POSTHOOK: Output: hdfs://### HDFS PATH ###
+0 0
+10 10
+100 200
+103 206
+104 208
+PREHOOK: query: explain vectorization detail
+SELECT key FROM src GROUP BY key ORDER BY key LIMIT 5
+PREHOOK: type: QUERY
+POSTHOOK: query: explain vectorization detail
+SELECT key FROM src GROUP BY key ORDER BY key LIMIT 5
+POSTHOOK: type: QUERY
+Plan optimized by CBO.
+
+Vertex dependency in root stage
+Reducer 2 <- Map 1 (SIMPLE_EDGE)
+Reducer 3 <- Reducer 2 (SIMPLE_EDGE)
+
+Stage-0
+ Fetch Operator
+ limit:5
+ Stage-1
+ Reducer 3 vectorized
+ File Output Operator [FS_19]
+ Limit [LIM_18] (rows=5 width=87)
+ Number of rows:5
+ Select Operator [SEL_17] (rows=250 width=87)
+ Output:["_col0"]
+ <-Reducer 2 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_16]
+ Group By Operator [GBY_15] (rows=250 width=87)
+ Output:["_col0"],keys:KEY._col0
+ <-Map 1 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_14]
+ PartitionCols:_col0
+ Group By Operator [GBY_13] (rows=250 width=87)
+ Output:["_col0"],keys:key
+ Top N Key Operator [TNK_12] (rows=500 width=87)
+ keys:key,sort order:+,top n:5
+ Select Operator [SEL_11] (rows=500 width=87)
+ Output:["key"]
+ TableScan [TS_0] (rows=500 width=87)
+ default@src,src,Tbl:COMPLETE,Col:COMPLETE,Output:["key"]
+
+PREHOOK: query: SELECT key FROM src GROUP BY key ORDER BY key LIMIT 5
+PREHOOK: type: QUERY
+PREHOOK: Input: default@src
+PREHOOK: Output: hdfs://### HDFS PATH ###
+POSTHOOK: query: SELECT key FROM src GROUP BY key ORDER BY key LIMIT 5
+POSTHOOK: type: QUERY
+POSTHOOK: Input: default@src
+POSTHOOK: Output: hdfs://### HDFS PATH ###
+0
+10
+100
+103
+104
+PREHOOK: query: explain vectorization detail
+SELECT src1.key, src2.value FROM src src1 JOIN src src2 ON (src1.key = src2.key) ORDER BY src1.key LIMIT 5
+PREHOOK: type: QUERY
+POSTHOOK: query: explain vectorization detail
+SELECT src1.key, src2.value FROM src src1 JOIN src src2 ON (src1.key = src2.key) ORDER BY src1.key LIMIT 5
+POSTHOOK: type: QUERY
+Plan optimized by CBO.
+
+Vertex dependency in root stage
+Reducer 2 <- Map 1 (SIMPLE_EDGE), Map 4 (SIMPLE_EDGE)
+Reducer 3 <- Reducer 2 (SIMPLE_EDGE)
+
+Stage-0
+ Fetch Operator
+ limit:5
+ Stage-1
+ Reducer 3 vectorized
+ File Output Operator [FS_37]
+ Limit [LIM_36] (rows=5 width=178)
+ Number of rows:5
+ Select Operator [SEL_35] (rows=791 width=178)
+ Output:["_col0","_col1"]
+ <-Reducer 2 [SIMPLE_EDGE]
+ SHUFFLE [RS_10]
+ Select Operator [SEL_9] (rows=791 width=178)
+ Output:["_col0","_col1"]
+ Merge Join Operator [MERGEJOIN_28] (rows=791 width=178)
+ Conds:RS_31._col0=RS_34._col0(Inner),Output:["_col0","_col2"]
+ <-Map 1 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_31]
+ PartitionCols:_col0
+ Select Operator [SEL_30] (rows=500 width=87)
+ Output:["_col0"]
+ Filter Operator [FIL_29] (rows=500 width=87)
+ predicate:key is not null
+ TableScan [TS_0] (rows=500 width=87)
+ default@src,src1,Tbl:COMPLETE,Col:COMPLETE,Output:["key"]
+ <-Map 4 [SIMPLE_EDGE] vectorized
+ SHUFFLE [RS_34]
+ PartitionCols:_col0
+ Select Operator [SEL_33] (rows=500 width=178)
+ Output:["_col0","_col1"]
+ Filter Operator [FIL_32] (rows=500 width=178)
+ predicate:key is not null
+ TableScan [TS_3] (rows=500 width=178)
+ default@src,src2,Tbl:COMPLETE,Col:COMPLETE,Output:["key","value"]
+
+PREHOOK: query: SELECT src1.key, src2.value FROM src src1 JOIN src src2 ON (src1.key = src2.key) ORDER BY src1.key LIMIT 5
+PREHOOK: type: QUERY
+PREHOOK: Input: default@src
+PREHOOK: Output: hdfs://### HDFS PATH ###
+POSTHOOK: query: SELECT src1.key, src2.value FROM src src1 JOIN src src2 ON (src1.key = src2.key) ORDER BY src1.key LIMIT 5
+POSTHOOK: type: QUERY
+POSTHOOK: Input: default@src
+POSTHOOK: Output: hdfs://### HDFS PATH ###
+0 val_0
+0 val_0
+0 val_0
+0 val_0
+0 val_0