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 2018/03/22 16:57:49 UTC
[26/34] hive git commit: HIVE-18979: Enable
AggregateReduceFunctionsRule from Calcite (Jesus Camacho Rodriguez,
reviewed by Ashutosh Chauhan)
http://git-wip-us.apache.org/repos/asf/hive/blob/5cb8867b/ql/src/test/results/clientpositive/llap/vectorization_0.q.out
----------------------------------------------------------------------
diff --git a/ql/src/test/results/clientpositive/llap/vectorization_0.q.out b/ql/src/test/results/clientpositive/llap/vectorization_0.q.out
index 1a846ab..b2db5a5 100644
--- a/ql/src/test/results/clientpositive/llap/vectorization_0.q.out
+++ b/ql/src/test/results/clientpositive/llap/vectorization_0.q.out
@@ -408,18 +408,18 @@ STAGE PLANS:
alias: alltypesorc
Statistics: Num rows: 12288 Data size: 36696 Basic stats: COMPLETE Column stats: COMPLETE
Select Operator
- expressions: ctinyint (type: tinyint)
- outputColumnNames: ctinyint
+ expressions: ctinyint (type: tinyint), UDFToDouble(ctinyint) (type: double), (UDFToDouble(ctinyint) * UDFToDouble(ctinyint)) (type: double)
+ outputColumnNames: _col0, _col1, _col2
Statistics: Num rows: 12288 Data size: 36696 Basic stats: COMPLETE Column stats: COMPLETE
Group By Operator
- aggregations: avg(ctinyint), variance(ctinyint), var_pop(ctinyint), var_samp(ctinyint), std(ctinyint), stddev(ctinyint), stddev_pop(ctinyint), stddev_samp(ctinyint)
+ aggregations: sum(_col0), count(_col0), sum(_col2), sum(_col1)
mode: hash
- outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7
- Statistics: Num rows: 1 Data size: 636 Basic stats: COMPLETE Column stats: COMPLETE
+ outputColumnNames: _col0, _col1, _col2, _col3
+ Statistics: Num rows: 1 Data size: 32 Basic stats: COMPLETE Column stats: COMPLETE
Reduce Output Operator
sort order:
- Statistics: Num rows: 1 Data size: 636 Basic stats: COMPLETE Column stats: COMPLETE
- value expressions: _col0 (type: struct<count:bigint,sum:double,input:tinyint>), _col1 (type: struct<count:bigint,sum:double,variance:double>), _col2 (type: struct<count:bigint,sum:double,variance:double>), _col3 (type: struct<count:bigint,sum:double,variance:double>), _col4 (type: struct<count:bigint,sum:double,variance:double>), _col5 (type: struct<count:bigint,sum:double,variance:double>), _col6 (type: struct<count:bigint,sum:double,variance:double>), _col7 (type: struct<count:bigint,sum:double,variance:double>)
+ Statistics: Num rows: 1 Data size: 32 Basic stats: COMPLETE Column stats: COMPLETE
+ value expressions: _col0 (type: bigint), _col1 (type: bigint), _col2 (type: double), _col3 (type: double)
Execution mode: vectorized, llap
LLAP IO: all inputs
Map Vectorization:
@@ -432,23 +432,28 @@ STAGE PLANS:
usesVectorUDFAdaptor: false
vectorized: true
Reducer 2
- Execution mode: llap
+ Execution mode: vectorized, llap
Reduce Vectorization:
enabled: true
enableConditionsMet: hive.vectorized.execution.reduce.enabled IS true, hive.execution.engine tez IN [tez, spark] IS true
- notVectorizedReason: GROUPBY operator: Vector aggregation : "variance" for input type: "STRUCT" and output type: "DOUBLE" and mode: FINAL not supported for evaluator GenericUDAFVarianceEvaluator
- vectorized: false
+ allNative: false
+ usesVectorUDFAdaptor: false
+ vectorized: true
Reduce Operator Tree:
Group By Operator
- aggregations: avg(VALUE._col0), variance(VALUE._col1), var_pop(VALUE._col2), var_samp(VALUE._col3), std(VALUE._col4), stddev(VALUE._col5), stddev_pop(VALUE._col6), stddev_samp(VALUE._col7)
+ aggregations: sum(VALUE._col0), count(VALUE._col1), sum(VALUE._col2), sum(VALUE._col3)
mode: mergepartial
- outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7
- Statistics: Num rows: 1 Data size: 64 Basic stats: COMPLETE Column stats: COMPLETE
- Reduce Output Operator
- key expressions: _col0 (type: double)
- sort order: +
+ outputColumnNames: _col0, _col1, _col2, _col3
+ Statistics: Num rows: 1 Data size: 32 Basic stats: COMPLETE Column stats: COMPLETE
+ Select Operator
+ expressions: (_col0 / _col1) (type: double), ((_col2 - ((_col3 * _col3) / _col1)) / _col1) (type: double), ((_col2 - ((_col3 * _col3) / _col1)) / CASE WHEN ((_col1 = 1L)) THEN (null) ELSE ((_col1 - 1)) END) (type: double), power(((_col2 - ((_col3 * _col3) / _col1)) / _col1), 0.5) (type: double), power(((_col2 - ((_col3 * _col3) / _col1)) / CASE WHEN ((_col1 = 1L)) THEN (null) ELSE ((_col1 - 1)) END), 0.5) (type: double)
+ outputColumnNames: _col0, _col1, _col3, _col4, _col7
Statistics: Num rows: 1 Data size: 64 Basic stats: COMPLETE Column stats: COMPLETE
- value expressions: _col1 (type: double), _col2 (type: double), _col3 (type: double), _col4 (type: double), _col5 (type: double), _col6 (type: double), _col7 (type: double)
+ Reduce Output Operator
+ key expressions: _col0 (type: double)
+ sort order: +
+ Statistics: Num rows: 1 Data size: 64 Basic stats: COMPLETE Column stats: COMPLETE
+ value expressions: _col1 (type: double), _col3 (type: double), _col4 (type: double), _col7 (type: double)
Reducer 3
Execution mode: vectorized, llap
Reduce Vectorization:
@@ -459,7 +464,7 @@ STAGE PLANS:
vectorized: true
Reduce Operator Tree:
Select Operator
- expressions: KEY.reducesinkkey0 (type: double), VALUE._col0 (type: double), VALUE._col1 (type: double), VALUE._col2 (type: double), VALUE._col3 (type: double), VALUE._col4 (type: double), VALUE._col5 (type: double), VALUE._col6 (type: double)
+ expressions: KEY.reducesinkkey0 (type: double), VALUE._col0 (type: double), VALUE._col0 (type: double), VALUE._col1 (type: double), VALUE._col2 (type: double), VALUE._col2 (type: double), VALUE._col2 (type: double), VALUE._col3 (type: double)
outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7
Statistics: Num rows: 1 Data size: 64 Basic stats: COMPLETE Column stats: COMPLETE
File Output Operator
@@ -504,7 +509,7 @@ ORDER BY c1
POSTHOOK: type: QUERY
POSTHOOK: Input: default@alltypesorc
#### A masked pattern was here ####
--4.344925324321378 1158.3003004768184 1158.3003004768184 1158.4265870337827 34.033811136527426 34.033811136527426 34.033811136527426 34.03566639620536
+-4.344925324321378 1158.3003004768175 1158.3003004768175 1158.426587033782 34.03381113652741 34.03381113652741 34.03381113652741 34.03566639620535
PREHOOK: query: EXPLAIN VECTORIZATION DETAIL
SELECT MIN(cbigint) as c1,
MAX(cbigint),
@@ -915,18 +920,18 @@ STAGE PLANS:
alias: alltypesorc
Statistics: Num rows: 12288 Data size: 73392 Basic stats: COMPLETE Column stats: COMPLETE
Select Operator
- expressions: cbigint (type: bigint)
- outputColumnNames: cbigint
+ expressions: cbigint (type: bigint), UDFToDouble(cbigint) (type: double), (UDFToDouble(cbigint) * UDFToDouble(cbigint)) (type: double)
+ outputColumnNames: _col0, _col1, _col2
Statistics: Num rows: 12288 Data size: 73392 Basic stats: COMPLETE Column stats: COMPLETE
Group By Operator
- aggregations: avg(cbigint), variance(cbigint), var_pop(cbigint), var_samp(cbigint), std(cbigint), stddev(cbigint), stddev_pop(cbigint), stddev_samp(cbigint)
+ aggregations: sum(_col0), count(_col0), sum(_col2), sum(_col1)
mode: hash
- outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7
- Statistics: Num rows: 1 Data size: 640 Basic stats: COMPLETE Column stats: COMPLETE
+ outputColumnNames: _col0, _col1, _col2, _col3
+ Statistics: Num rows: 1 Data size: 32 Basic stats: COMPLETE Column stats: COMPLETE
Reduce Output Operator
sort order:
- Statistics: Num rows: 1 Data size: 640 Basic stats: COMPLETE Column stats: COMPLETE
- value expressions: _col0 (type: struct<count:bigint,sum:double,input:bigint>), _col1 (type: struct<count:bigint,sum:double,variance:double>), _col2 (type: struct<count:bigint,sum:double,variance:double>), _col3 (type: struct<count:bigint,sum:double,variance:double>), _col4 (type: struct<count:bigint,sum:double,variance:double>), _col5 (type: struct<count:bigint,sum:double,variance:double>), _col6 (type: struct<count:bigint,sum:double,variance:double>), _col7 (type: struct<count:bigint,sum:double,variance:double>)
+ Statistics: Num rows: 1 Data size: 32 Basic stats: COMPLETE Column stats: COMPLETE
+ value expressions: _col0 (type: bigint), _col1 (type: bigint), _col2 (type: double), _col3 (type: double)
Execution mode: vectorized, llap
LLAP IO: all inputs
Map Vectorization:
@@ -939,23 +944,28 @@ STAGE PLANS:
usesVectorUDFAdaptor: false
vectorized: true
Reducer 2
- Execution mode: llap
+ Execution mode: vectorized, llap
Reduce Vectorization:
enabled: true
enableConditionsMet: hive.vectorized.execution.reduce.enabled IS true, hive.execution.engine tez IN [tez, spark] IS true
- notVectorizedReason: GROUPBY operator: Vector aggregation : "variance" for input type: "STRUCT" and output type: "DOUBLE" and mode: FINAL not supported for evaluator GenericUDAFVarianceEvaluator
- vectorized: false
+ allNative: false
+ usesVectorUDFAdaptor: false
+ vectorized: true
Reduce Operator Tree:
Group By Operator
- aggregations: avg(VALUE._col0), variance(VALUE._col1), var_pop(VALUE._col2), var_samp(VALUE._col3), std(VALUE._col4), stddev(VALUE._col5), stddev_pop(VALUE._col6), stddev_samp(VALUE._col7)
+ aggregations: sum(VALUE._col0), count(VALUE._col1), sum(VALUE._col2), sum(VALUE._col3)
mode: mergepartial
- outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7
- Statistics: Num rows: 1 Data size: 64 Basic stats: COMPLETE Column stats: COMPLETE
- Reduce Output Operator
- key expressions: _col0 (type: double)
- sort order: +
+ outputColumnNames: _col0, _col1, _col2, _col3
+ Statistics: Num rows: 1 Data size: 32 Basic stats: COMPLETE Column stats: COMPLETE
+ Select Operator
+ expressions: (_col0 / _col1) (type: double), ((_col2 - ((_col3 * _col3) / _col1)) / _col1) (type: double), ((_col2 - ((_col3 * _col3) / _col1)) / CASE WHEN ((_col1 = 1L)) THEN (null) ELSE ((_col1 - 1)) END) (type: double), power(((_col2 - ((_col3 * _col3) / _col1)) / _col1), 0.5) (type: double), power(((_col2 - ((_col3 * _col3) / _col1)) / CASE WHEN ((_col1 = 1L)) THEN (null) ELSE ((_col1 - 1)) END), 0.5) (type: double)
+ outputColumnNames: _col0, _col1, _col3, _col4, _col7
Statistics: Num rows: 1 Data size: 64 Basic stats: COMPLETE Column stats: COMPLETE
- value expressions: _col1 (type: double), _col2 (type: double), _col3 (type: double), _col4 (type: double), _col5 (type: double), _col6 (type: double), _col7 (type: double)
+ Reduce Output Operator
+ key expressions: _col0 (type: double)
+ sort order: +
+ Statistics: Num rows: 1 Data size: 64 Basic stats: COMPLETE Column stats: COMPLETE
+ value expressions: _col1 (type: double), _col3 (type: double), _col4 (type: double), _col7 (type: double)
Reducer 3
Execution mode: vectorized, llap
Reduce Vectorization:
@@ -966,7 +976,7 @@ STAGE PLANS:
vectorized: true
Reduce Operator Tree:
Select Operator
- expressions: KEY.reducesinkkey0 (type: double), VALUE._col0 (type: double), VALUE._col1 (type: double), VALUE._col2 (type: double), VALUE._col3 (type: double), VALUE._col4 (type: double), VALUE._col5 (type: double), VALUE._col6 (type: double)
+ expressions: KEY.reducesinkkey0 (type: double), VALUE._col0 (type: double), VALUE._col0 (type: double), VALUE._col1 (type: double), VALUE._col2 (type: double), VALUE._col2 (type: double), VALUE._col2 (type: double), VALUE._col3 (type: double)
outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7
Statistics: Num rows: 1 Data size: 64 Basic stats: COMPLETE Column stats: COMPLETE
File Output Operator
@@ -1011,7 +1021,7 @@ ORDER BY c1
POSTHOOK: type: QUERY
POSTHOOK: Input: default@alltypesorc
#### A masked pattern was here ####
--1.8515862077935246E8 2.07689300543081907E18 2.07689300543081907E18 2.07711944383088768E18 1.441142951074188E9 1.441142951074188E9 1.441142951074188E9 1.4412215110214279E9
+-1.8515862077935246E8 2.07689300543070106E18 2.07689300543070106E18 2.07711944383076992E18 1.441142951074147E9 1.441142951074147E9 1.441142951074147E9 1.441221511021387E9
PREHOOK: query: EXPLAIN VECTORIZATION DETAIL
SELECT MIN(cfloat) as c1,
MAX(cfloat),
@@ -1422,18 +1432,18 @@ STAGE PLANS:
alias: alltypesorc
Statistics: Num rows: 12288 Data size: 36696 Basic stats: COMPLETE Column stats: COMPLETE
Select Operator
- expressions: cfloat (type: float)
- outputColumnNames: cfloat
+ expressions: cfloat (type: float), UDFToDouble(cfloat) (type: double), (UDFToDouble(cfloat) * UDFToDouble(cfloat)) (type: double)
+ outputColumnNames: _col0, _col1, _col2
Statistics: Num rows: 12288 Data size: 36696 Basic stats: COMPLETE Column stats: COMPLETE
Group By Operator
- aggregations: avg(cfloat), variance(cfloat), var_pop(cfloat), var_samp(cfloat), std(cfloat), stddev(cfloat), stddev_pop(cfloat), stddev_samp(cfloat)
+ aggregations: sum(_col0), count(_col0), sum(_col2), sum(_col1)
mode: hash
- outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7
- Statistics: Num rows: 1 Data size: 636 Basic stats: COMPLETE Column stats: COMPLETE
+ outputColumnNames: _col0, _col1, _col2, _col3
+ Statistics: Num rows: 1 Data size: 32 Basic stats: COMPLETE Column stats: COMPLETE
Reduce Output Operator
sort order:
- Statistics: Num rows: 1 Data size: 636 Basic stats: COMPLETE Column stats: COMPLETE
- value expressions: _col0 (type: struct<count:bigint,sum:double,input:float>), _col1 (type: struct<count:bigint,sum:double,variance:double>), _col2 (type: struct<count:bigint,sum:double,variance:double>), _col3 (type: struct<count:bigint,sum:double,variance:double>), _col4 (type: struct<count:bigint,sum:double,variance:double>), _col5 (type: struct<count:bigint,sum:double,variance:double>), _col6 (type: struct<count:bigint,sum:double,variance:double>), _col7 (type: struct<count:bigint,sum:double,variance:double>)
+ Statistics: Num rows: 1 Data size: 32 Basic stats: COMPLETE Column stats: COMPLETE
+ value expressions: _col0 (type: double), _col1 (type: bigint), _col2 (type: double), _col3 (type: double)
Execution mode: vectorized, llap
LLAP IO: all inputs
Map Vectorization:
@@ -1446,23 +1456,28 @@ STAGE PLANS:
usesVectorUDFAdaptor: false
vectorized: true
Reducer 2
- Execution mode: llap
+ Execution mode: vectorized, llap
Reduce Vectorization:
enabled: true
enableConditionsMet: hive.vectorized.execution.reduce.enabled IS true, hive.execution.engine tez IN [tez, spark] IS true
- notVectorizedReason: GROUPBY operator: Vector aggregation : "variance" for input type: "STRUCT" and output type: "DOUBLE" and mode: FINAL not supported for evaluator GenericUDAFVarianceEvaluator
- vectorized: false
+ allNative: false
+ usesVectorUDFAdaptor: false
+ vectorized: true
Reduce Operator Tree:
Group By Operator
- aggregations: avg(VALUE._col0), variance(VALUE._col1), var_pop(VALUE._col2), var_samp(VALUE._col3), std(VALUE._col4), stddev(VALUE._col5), stddev_pop(VALUE._col6), stddev_samp(VALUE._col7)
+ aggregations: sum(VALUE._col0), count(VALUE._col1), sum(VALUE._col2), sum(VALUE._col3)
mode: mergepartial
- outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7
- Statistics: Num rows: 1 Data size: 64 Basic stats: COMPLETE Column stats: COMPLETE
- Reduce Output Operator
- key expressions: _col0 (type: double)
- sort order: +
+ outputColumnNames: _col0, _col1, _col2, _col3
+ Statistics: Num rows: 1 Data size: 32 Basic stats: COMPLETE Column stats: COMPLETE
+ Select Operator
+ expressions: (_col0 / _col1) (type: double), ((_col2 - ((_col3 * _col3) / _col1)) / _col1) (type: double), ((_col2 - ((_col3 * _col3) / _col1)) / CASE WHEN ((_col1 = 1L)) THEN (null) ELSE ((_col1 - 1)) END) (type: double), power(((_col2 - ((_col3 * _col3) / _col1)) / _col1), 0.5) (type: double), power(((_col2 - ((_col3 * _col3) / _col1)) / CASE WHEN ((_col1 = 1L)) THEN (null) ELSE ((_col1 - 1)) END), 0.5) (type: double)
+ outputColumnNames: _col0, _col1, _col3, _col4, _col7
Statistics: Num rows: 1 Data size: 64 Basic stats: COMPLETE Column stats: COMPLETE
- value expressions: _col1 (type: double), _col2 (type: double), _col3 (type: double), _col4 (type: double), _col5 (type: double), _col6 (type: double), _col7 (type: double)
+ Reduce Output Operator
+ key expressions: _col0 (type: double)
+ sort order: +
+ Statistics: Num rows: 1 Data size: 64 Basic stats: COMPLETE Column stats: COMPLETE
+ value expressions: _col1 (type: double), _col3 (type: double), _col4 (type: double), _col7 (type: double)
Reducer 3
Execution mode: vectorized, llap
Reduce Vectorization:
@@ -1473,7 +1488,7 @@ STAGE PLANS:
vectorized: true
Reduce Operator Tree:
Select Operator
- expressions: KEY.reducesinkkey0 (type: double), VALUE._col0 (type: double), VALUE._col1 (type: double), VALUE._col2 (type: double), VALUE._col3 (type: double), VALUE._col4 (type: double), VALUE._col5 (type: double), VALUE._col6 (type: double)
+ expressions: KEY.reducesinkkey0 (type: double), VALUE._col0 (type: double), VALUE._col0 (type: double), VALUE._col1 (type: double), VALUE._col2 (type: double), VALUE._col2 (type: double), VALUE._col2 (type: double), VALUE._col3 (type: double)
outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7
Statistics: Num rows: 1 Data size: 64 Basic stats: COMPLETE Column stats: COMPLETE
File Output Operator
@@ -1518,7 +1533,7 @@ ORDER BY c1
POSTHOOK: type: QUERY
POSTHOOK: Input: default@alltypesorc
#### A masked pattern was here ####
--4.303895780321011 1163.8972588604984 1163.8972588604984 1164.0241556397025 34.115938487171924 34.115938487171924 34.115938487171924 34.11779822379666
+-4.303895780321011 1163.8972588605056 1163.8972588605056 1164.0241556397098 34.11593848717203 34.11593848717203 34.11593848717203 34.11779822379677
WARNING: Comparing a bigint and a double may result in a loss of precision.
PREHOOK: query: EXPLAIN VECTORIZATION DETAIL
SELECT AVG(cbigint),
@@ -1606,25 +1621,26 @@ STAGE PLANS:
predicate: (((UDFToShort(ctinyint) >= csmallint) and (cboolean2 = 1) and (UDFToInteger(ctinyint) = 3569)) or (CAST( cint AS decimal(13,3)) <> 79.553) or (UDFToDouble(cbigint) < cdouble) or (cstring2 like '%b%')) (type: boolean)
Statistics: Num rows: 12288 Data size: 1210980 Basic stats: COMPLETE Column stats: COMPLETE
Select Operator
- expressions: ctinyint (type: tinyint), cbigint (type: bigint), cfloat (type: float)
- outputColumnNames: ctinyint, cbigint, cfloat
+ expressions: cbigint (type: bigint), cfloat (type: float), ctinyint (type: tinyint), UDFToDouble(cbigint) (type: double), (UDFToDouble(cbigint) * UDFToDouble(cbigint)) (type: double)
+ outputColumnNames: _col0, _col1, _col2, _col3, _col4
Select Vectorization:
className: VectorSelectOperator
native: true
- projectedOutputColumnNums: [0, 3, 4]
+ projectedOutputColumnNums: [3, 4, 0, 14, 17]
+ selectExpressions: CastLongToDouble(col 3:bigint) -> 14:double, DoubleColMultiplyDoubleColumn(col 15:double, col 16:double)(children: CastLongToDouble(col 3:bigint) -> 15:double, CastLongToDouble(col 3:bigint) -> 16:double) -> 17:double
Statistics: Num rows: 12288 Data size: 1210980 Basic stats: COMPLETE Column stats: COMPLETE
Group By Operator
- aggregations: avg(cbigint), stddev_pop(cbigint), var_samp(cbigint), count(), sum(cfloat), min(ctinyint)
+ aggregations: sum(_col0), count(_col0), sum(_col4), sum(_col3), count(), sum(_col1), min(_col2)
Group By Vectorization:
- aggregators: VectorUDAFAvgLong(col 3:bigint) -> struct<count:bigint,sum:double,input:bigint>, VectorUDAFVarLong(col 3:bigint) -> struct<count:bigint,sum:double,variance:double> aggregation: stddev_pop, VectorUDAFVarLong(col 3:bigint) -> struct<count:bigint,sum:double,variance:double> aggregation: var_samp, VectorUDAFCountStar(*) -> bigint, VectorUDAFSumDouble(col 4:float) -> double, VectorUDAFMinLong(col 0:tinyint) -> tinyint
+ aggregators: VectorUDAFSumLong(col 3:bigint) -> bigint, VectorUDAFCount(col 3:bigint) -> bigint, VectorUDAFSumDouble(col 17:double) -> double, VectorUDAFSumDouble(col 14:double) -> double, VectorUDAFCountStar(*) -> bigint, VectorUDAFSumDouble(col 4:float) -> double, VectorUDAFMinLong(col 0:tinyint) -> tinyint
className: VectorGroupByOperator
groupByMode: HASH
native: false
vectorProcessingMode: HASH
- projectedOutputColumnNums: [0, 1, 2, 3, 4, 5]
+ projectedOutputColumnNums: [0, 1, 2, 3, 4, 5, 6]
mode: hash
- outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5
- Statistics: Num rows: 1 Data size: 260 Basic stats: COMPLETE Column stats: COMPLETE
+ outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6
+ Statistics: Num rows: 1 Data size: 52 Basic stats: COMPLETE Column stats: COMPLETE
Reduce Output Operator
sort order:
Reduce Sink Vectorization:
@@ -1632,9 +1648,9 @@ STAGE PLANS:
keyColumnNums: []
native: true
nativeConditionsMet: hive.vectorized.execution.reducesink.new.enabled IS true, hive.execution.engine tez IN [tez, spark] IS true, No PTF TopN IS true, No DISTINCT columns IS true, BinarySortableSerDe for keys IS true, LazyBinarySerDe for values IS true
- valueColumnNums: [0, 1, 2, 3, 4, 5]
- Statistics: Num rows: 1 Data size: 260 Basic stats: COMPLETE Column stats: COMPLETE
- value expressions: _col0 (type: struct<count:bigint,sum:double,input:bigint>), _col1 (type: struct<count:bigint,sum:double,variance:double>), _col2 (type: struct<count:bigint,sum:double,variance:double>), _col3 (type: bigint), _col4 (type: double), _col5 (type: tinyint)
+ valueColumnNums: [0, 1, 2, 3, 4, 5, 6]
+ Statistics: Num rows: 1 Data size: 52 Basic stats: COMPLETE Column stats: COMPLETE
+ value expressions: _col0 (type: bigint), _col1 (type: bigint), _col2 (type: double), _col3 (type: double), _col4 (type: bigint), _col5 (type: double), _col6 (type: tinyint)
Execution mode: vectorized, llap
LLAP IO: all inputs
Map Vectorization:
@@ -1651,26 +1667,49 @@ STAGE PLANS:
includeColumns: [0, 1, 2, 3, 4, 5, 7, 11]
dataColumns: ctinyint:tinyint, csmallint:smallint, cint:int, cbigint:bigint, cfloat:float, cdouble:double, cstring1:string, cstring2:string, ctimestamp1:timestamp, ctimestamp2:timestamp, cboolean1:boolean, cboolean2:boolean
partitionColumnCount: 0
- scratchColumnTypeNames: [decimal(13,3), double]
+ scratchColumnTypeNames: [decimal(13,3), double, double, double, double]
Reducer 2
- Execution mode: llap
+ Execution mode: vectorized, llap
Reduce Vectorization:
enabled: true
enableConditionsMet: hive.vectorized.execution.reduce.enabled IS true, hive.execution.engine tez IN [tez, spark] IS true
- notVectorizedReason: GROUPBY operator: Vector aggregation : "stddev_pop" for input type: "STRUCT" and output type: "DOUBLE" and mode: FINAL not supported for evaluator GenericUDAFStdEvaluator
- vectorized: false
+ reduceColumnNullOrder:
+ reduceColumnSortOrder:
+ allNative: false
+ usesVectorUDFAdaptor: false
+ vectorized: true
+ rowBatchContext:
+ dataColumnCount: 7
+ dataColumns: VALUE._col0:bigint, VALUE._col1:bigint, VALUE._col2:double, VALUE._col3:double, VALUE._col4:bigint, VALUE._col5:double, VALUE._col6:tinyint
+ partitionColumnCount: 0
+ scratchColumnTypeNames: []
Reduce Operator Tree:
Group By Operator
- aggregations: avg(VALUE._col0), stddev_pop(VALUE._col1), var_samp(VALUE._col2), count(VALUE._col3), sum(VALUE._col4), min(VALUE._col5)
+ aggregations: sum(VALUE._col0), count(VALUE._col1), sum(VALUE._col2), sum(VALUE._col3), count(VALUE._col4), sum(VALUE._col5), min(VALUE._col6)
+ Group By Vectorization:
+ aggregators: VectorUDAFSumLong(col 0:bigint) -> bigint, VectorUDAFCountMerge(col 1:bigint) -> bigint, VectorUDAFSumDouble(col 2:double) -> double, VectorUDAFSumDouble(col 3:double) -> double, VectorUDAFCountMerge(col 4:bigint) -> bigint, VectorUDAFSumDouble(col 5:double) -> double, VectorUDAFMinLong(col 6:tinyint) -> tinyint
+ className: VectorGroupByOperator
+ groupByMode: MERGEPARTIAL
+ native: false
+ vectorProcessingMode: GLOBAL
+ projectedOutputColumnNums: [0, 1, 2, 3, 4, 5, 6]
mode: mergepartial
- outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5
- Statistics: Num rows: 1 Data size: 44 Basic stats: COMPLETE Column stats: COMPLETE
+ outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6
+ Statistics: Num rows: 1 Data size: 52 Basic stats: COMPLETE Column stats: COMPLETE
Select Operator
- expressions: _col0 (type: double), (- _col0) (type: double), (-6432.0D + _col0) (type: double), _col1 (type: double), (- (-6432.0D + _col0)) (type: double), ((- (-6432.0D + _col0)) + (-6432.0D + _col0)) (type: double), _col2 (type: double), (- (-6432.0D + _col0)) (type: double), (-6432.0D + (- (-6432.0D + _col0))) (type: double), (- (-6432.0D + _col0)) (type: double), ((- (-6432.0D + _col0)) / (- (-6432.0D + _col0))) (type: double), _col3 (type: bigint), _col4 (type: double), (_col2 % _col1) (type: double), (- _col2) (type: double), ((- (-6432.0D + _col0)) * (- _col0)) (type: double), _col5 (type: tinyint), (- _col5) (type: tinyint)
+ expressions: (_col0 / _col1) (type: double), (- (_col0 / _col1)) (type: double), (-6432.0D + (_col0 / _col1)) (type: double), power(((_col2 - ((_col3 * _col3) / _col1)) / _col1), 0.5) (type: double), (- (-6432.0D + (_col0 / _col1))) (type: double), ((- (-6432.0D + (_col0 / _col1))) + (-6432.0D + (_col0 / _col1))) (type: double), ((_col2 - ((_col3 * _col3) / _col1)) / CASE WHEN ((_col1 = 1L)) THEN (null) ELSE ((_col1 - 1)) END) (type: double), (- (-6432.0D + (_col0 / _col1))) (type: double), (-6432.0D + (- (-6432.0D + (_col0 / _col1)))) (type: double), (- (-6432.0D + (_col0 / _col1))) (type: double), ((- (-6432.0D + (_col0 / _col1))) / (- (-6432.0D + (_col0 / _col1)))) (type: double), _col4 (type: bigint), _col5 (type: double), (((_col2 - ((_col3 * _col3) / _col1)) / CASE WHEN ((_col1 = 1L)) THEN (null) ELSE ((_col1 - 1)) END) % power(((_col2 - ((_col3 * _col3) / _col1)) / _col1), 0.5)) (type: double), (- ((_col2 - ((_col3 * _col3) / _col1)) / CASE WHEN ((_col1 = 1L
)) THEN (null) ELSE ((_col1 - 1)) END)) (type: double), ((- (-6432.0D + (_col0 / _col1))) * (- (_col0 / _col1))) (type: double), _col6 (type: tinyint), (- _col6) (type: tinyint)
outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8, _col9, _col10, _col11, _col12, _col13, _col14, _col15, _col16, _col17
+ Select Vectorization:
+ className: VectorSelectOperator
+ native: true
+ projectedOutputColumnNums: [7, 9, 10, 8, 11, 13, 14, 12, 19, 18, 22, 4, 5, 25, 20, 28, 6, 27]
+ selectExpressions: LongColDivideLongColumn(col 0:bigint, col 1:bigint) -> 7:double, DoubleColUnaryMinus(col 8:double)(children: LongColDivideLongColumn(col 0:bigint, col 1:bigint) -> 8:double) -> 9:double, DoubleScalarAddDoubleColumn(val -6432.0, col 8:double)(children: LongColDivideLongColumn(col 0:bigint, col 1:bigint) -> 8:double) -> 10:double, FuncPowerDoubleToDouble(col 11:double)(children: DoubleColDivideLongColumn(col 8:double, col 1:bigint)(children: DoubleColSubtractDoubleColumn(col 2:double, col 11:double)(children: DoubleColDivideLongColumn(col 8:double, col 1:bigint)(children: DoubleColMultiplyDoubleColumn(col 3:double, col 3:double) -> 8:double) -> 11:double) -> 8:double) -> 11:double) -> 8:double, DoubleColUnaryMinus(col 12:double)(children: DoubleScalarAddDoubleColumn(val -6432.0, col 11:double)(children: LongColDivideLongColumn(col 0:bigint, col 1:bigint) -> 11:double) -> 12:double) -> 11:double, DoubleColAddDoubleColumn(col 12:double, col 14:do
uble)(children: DoubleColUnaryMinus(col 13:double)(children: DoubleScalarAddDoubleColumn(val -6432.0, col 12:double)(children: LongColDivideLongColumn(col 0:bigint, col 1:bigint) -> 12:double) -> 13:double) -> 12:double, DoubleScalarAddDoubleColumn(val -6432.0, col 13:double)(children: LongColDivideLongColumn(col 0:bigint, col 1:bigint) -> 13:double) -> 14:double) -> 13:double, DoubleColDivideLongColumn(col 12:double, col 17:bigint)(children: DoubleColSubtractDoubleColumn(col 2:double, col 14:double)(children: DoubleColDivideLongColumn(col 12:double, col 1:bigint)(children: DoubleColMultiplyDoubleColumn(col 3:double, col 3:double) -> 12:double) -> 14:double) -> 12:double, IfExprNullCondExpr(col 15:boolean, null, col 16:bigint)(children: LongColEqualLongScalar(col 1:bigint, val 1) -> 15:boolean, LongColSubtractLongScalar(col 1:bigint, val 1) -> 16:bigint) -> 17:bigint) -> 14:double, DoubleColUnaryMinus(col 18:double)(children: DoubleScalarAddDoubleColumn(val -6432.0, col 12:double)(c
hildren: LongColDivideLongColumn(col 0:bigint, col 1:bigint) -> 12:double) -> 18:double) -> 12:double, DoubleScalarAddDoubleColumn(val -6432.0, col 18:double)(children: DoubleColUnaryMinus(col 19:double)(children: DoubleScalarAddDoubleColumn(val -6432.0, col 18:double)(children: LongColDivideLongColumn(col 0:bigint, col 1:bigint) -> 18:double) -> 19:double) -> 18:double) -> 19:double, DoubleColUnaryMinus(col 20:double)(children: DoubleScalarAddDoubleColumn(val -6432.0, col 18:double)(children: LongColDivideLongColumn(col 0:bigint, col 1:bigint) -> 18:double) -> 20:double) -> 18:double, DoubleColDivideDoubleColumn(col 20:double, col 21:double)(children: DoubleColUnaryMinus(col 21:double)(children: DoubleScalarAddDoubleColumn(val -6432.0, col 20:double)(children: LongColDivideLongColumn(col 0:bigint, col 1:bigint) -> 20:double) -> 21:double) -> 20:double, DoubleColUnaryMinus(col 22:double)(children: DoubleScalarAddDoubleColumn(val -6432.0, col 21:double)(children: LongColDivideLongCol
umn(col 0:bigint, col 1:bigint) -> 21:double) -> 22:double) -> 21:double) -> 22:double, DoubleColModuloDoubleColumn(col 21:double, col 20:double)(children: DoubleColDivideLongColumn(col 20:double, col 24:bigint)(children: DoubleColSubtractDoubleColumn(col 2:double, col 21:double)(children: DoubleColDivideLongColumn(col 20:double, col 1:bigint)(children: DoubleColMultiplyDoubleColumn(col 3:double, col 3:double) -> 20:double) -> 21:double) -> 20:double, IfExprNullCondExpr(col 17:boolean, null, col 23:bigint)(children: LongColEqualLongScalar(col 1:bigint, val 1) -> 17:boolean, LongColSubtractLongScalar(col 1:bigint, val 1) -> 23:bigint) -> 24:bigint) -> 21:double, FuncPowerDoubleToDouble(col 25:double)(children: DoubleColDivideLongColumn(col 20:double, col 1:bigint)(children: DoubleColSubtractDoubleColumn(col 2:double, col 25:double)(children: DoubleColDivideLongColumn(col 20:double, col 1:bigint)(children: DoubleColMultiplyDoubleColumn(col 3:double, col 3:double) -> 20:double) -> 25:d
ouble) -> 20:double) -> 25:double) -> 20:double) -> 25:double, DoubleColUnaryMinus(col 21:double)(children: DoubleColDivideLongColumn(col 20:double, col 27:bigint)(children: DoubleColSubtractDoubleColumn(col 2:double, col 21:double)(children: DoubleColDivideLongColumn(col 20:double, col 1:bigint)(children: DoubleColMultiplyDoubleColumn(col 3:double, col 3:double) -> 20:double) -> 21:double) -> 20:double, IfExprNullCondExpr(col 24:boolean, null, col 26:bigint)(children: LongColEqualLongScalar(col 1:bigint, val 1) -> 24:boolean, LongColSubtractLongScalar(col 1:bigint, val 1) -> 26:bigint) -> 27:bigint) -> 21:double) -> 20:double, DoubleColMultiplyDoubleColumn(col 21:double, col 29:double)(children: DoubleColUnaryMinus(col 28:double)(children: DoubleScalarAddDoubleColumn(val -6432.0, col 21:double)(children: LongColDivideLongColumn(col 0:bigint, col 1:bigint) -> 21:double) -> 28:double) -> 21:double, DoubleColUnaryMinus(col 28:double)(children: LongColDivideLongColumn(col 0:bigint, col
1:bigint) -> 28:double) -> 29:double) -> 28:double, LongColUnaryMinus(col 6:tinyint) -> 27:tinyint
Statistics: Num rows: 1 Data size: 136 Basic stats: COMPLETE Column stats: COMPLETE
File Output Operator
compressed: false
+ File Sink Vectorization:
+ className: VectorFileSinkOperator
+ native: false
Statistics: Num rows: 1 Data size: 136 Basic stats: COMPLETE Column stats: COMPLETE
table:
input format: org.apache.hadoop.mapred.SequenceFileInputFormat
@@ -1740,7 +1779,7 @@ WHERE (((cstring2 LIKE '%b%')
POSTHOOK: type: QUERY
POSTHOOK: Input: default@alltypesorc
#### A masked pattern was here ####
--3.875652215945533E8 3.875652215945533E8 -3.875716535945533E8 1.436387455459401E9 3.875716535945533E8 0.0 2.06347151720204902E18 3.875716535945533E8 3.875652215945533E8 3.875716535945533E8 1.0 10934 -37224.52399241924 1.0517370547117279E9 -2.06347151720204902E18 1.5020929380914048E17 -64 64
+-3.875652215945533E8 3.875652215945533E8 -3.875716535945533E8 1.4363874554593627E9 3.875716535945533E8 0.0 2.0634715172019392E18 3.875716535945533E8 3.875652215945533E8 3.875716535945533E8 1.0 10934 -37224.52399241924 1.0516820315185745E9 -2.0634715172019392E18 1.5020929380914048E17 -64 64
PREHOOK: query: EXPLAIN extended
select count(*) from alltypesorc
where (((cstring1 LIKE 'a%') or ((cstring1 like 'b%') or (cstring1 like 'c%'))) or
http://git-wip-us.apache.org/repos/asf/hive/blob/5cb8867b/ql/src/test/results/clientpositive/llap/vectorization_1.q.out
----------------------------------------------------------------------
diff --git a/ql/src/test/results/clientpositive/llap/vectorization_1.q.out b/ql/src/test/results/clientpositive/llap/vectorization_1.q.out
index dbee077..c87926c 100644
--- a/ql/src/test/results/clientpositive/llap/vectorization_1.q.out
+++ b/ql/src/test/results/clientpositive/llap/vectorization_1.q.out
@@ -72,25 +72,26 @@ STAGE PLANS:
predicate: (((cdouble > UDFToDouble(ctinyint)) and (cboolean2 > 0)) or (UDFToLong(cint) > cbigint) or (cbigint < UDFToLong(ctinyint)) or (cboolean1 < 0)) (type: boolean)
Statistics: Num rows: 12288 Data size: 330276 Basic stats: COMPLETE Column stats: COMPLETE
Select Operator
- expressions: ctinyint (type: tinyint), cint (type: int), cfloat (type: float), cdouble (type: double)
- outputColumnNames: ctinyint, cint, cfloat, cdouble
+ expressions: ctinyint (type: tinyint), cfloat (type: float), cint (type: int), cdouble (type: double), UDFToDouble(ctinyint) (type: double), (UDFToDouble(ctinyint) * UDFToDouble(ctinyint)) (type: double), (cdouble * cdouble) (type: double)
+ outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6
Select Vectorization:
className: VectorSelectOperator
native: true
- projectedOutputColumnNums: [0, 2, 4, 5]
+ projectedOutputColumnNums: [0, 4, 2, 5, 13, 16, 14]
+ selectExpressions: CastLongToDouble(col 0:tinyint) -> 13:double, DoubleColMultiplyDoubleColumn(col 14:double, col 15:double)(children: CastLongToDouble(col 0:tinyint) -> 14:double, CastLongToDouble(col 0:tinyint) -> 15:double) -> 16:double, DoubleColMultiplyDoubleColumn(col 5:double, col 5:double) -> 14:double
Statistics: Num rows: 12288 Data size: 330276 Basic stats: COMPLETE Column stats: COMPLETE
Group By Operator
- aggregations: var_pop(ctinyint), sum(cfloat), max(ctinyint), max(cint), var_samp(cdouble), count(cint)
+ aggregations: sum(_col5), sum(_col4), count(_col0), sum(_col1), max(_col0), max(_col2), sum(_col6), sum(_col3), count(_col3), count(_col2)
Group By Vectorization:
- aggregators: VectorUDAFVarLong(col 0:tinyint) -> struct<count:bigint,sum:double,variance:double> aggregation: var_pop, VectorUDAFSumDouble(col 4:float) -> double, VectorUDAFMaxLong(col 0:tinyint) -> tinyint, VectorUDAFMaxLong(col 2:int) -> int, VectorUDAFVarDouble(col 5:double) -> struct<count:bigint,sum:double,variance:double> aggregation: var_samp, VectorUDAFCount(col 2:int) -> bigint
+ aggregators: VectorUDAFSumDouble(col 16:double) -> double, VectorUDAFSumDouble(col 13:double) -> double, VectorUDAFCount(col 0:tinyint) -> bigint, VectorUDAFSumDouble(col 4:float) -> double, VectorUDAFMaxLong(col 0:tinyint) -> tinyint, VectorUDAFMaxLong(col 2:int) -> int, VectorUDAFSumDouble(col 14:double) -> double, VectorUDAFSumDouble(col 5:double) -> double, VectorUDAFCount(col 5:double) -> bigint, VectorUDAFCount(col 2:int) -> bigint
className: VectorGroupByOperator
groupByMode: HASH
native: false
vectorProcessingMode: HASH
- projectedOutputColumnNums: [0, 1, 2, 3, 4, 5]
+ projectedOutputColumnNums: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
mode: hash
- outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5
- Statistics: Num rows: 1 Data size: 184 Basic stats: COMPLETE Column stats: COMPLETE
+ outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8, _col9
+ Statistics: Num rows: 1 Data size: 72 Basic stats: COMPLETE Column stats: COMPLETE
Reduce Output Operator
sort order:
Reduce Sink Vectorization:
@@ -98,9 +99,9 @@ STAGE PLANS:
keyColumnNums: []
native: true
nativeConditionsMet: hive.vectorized.execution.reducesink.new.enabled IS true, hive.execution.engine tez IN [tez, spark] IS true, No PTF TopN IS true, No DISTINCT columns IS true, BinarySortableSerDe for keys IS true, LazyBinarySerDe for values IS true
- valueColumnNums: [0, 1, 2, 3, 4, 5]
- Statistics: Num rows: 1 Data size: 184 Basic stats: COMPLETE Column stats: COMPLETE
- value expressions: _col0 (type: struct<count:bigint,sum:double,variance:double>), _col1 (type: double), _col2 (type: tinyint), _col3 (type: int), _col4 (type: struct<count:bigint,sum:double,variance:double>), _col5 (type: bigint)
+ valueColumnNums: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
+ Statistics: Num rows: 1 Data size: 72 Basic stats: COMPLETE Column stats: COMPLETE
+ value expressions: _col0 (type: double), _col1 (type: double), _col2 (type: bigint), _col3 (type: double), _col4 (type: tinyint), _col5 (type: int), _col6 (type: double), _col7 (type: double), _col8 (type: bigint), _col9 (type: bigint)
Execution mode: vectorized, llap
LLAP IO: all inputs
Map Vectorization:
@@ -117,26 +118,49 @@ STAGE PLANS:
includeColumns: [0, 2, 3, 4, 5, 10, 11]
dataColumns: ctinyint:tinyint, csmallint:smallint, cint:int, cbigint:bigint, cfloat:float, cdouble:double, cstring1:string, cstring2:string, ctimestamp1:timestamp, ctimestamp2:timestamp, cboolean1:boolean, cboolean2:boolean
partitionColumnCount: 0
- scratchColumnTypeNames: [double]
+ scratchColumnTypeNames: [double, double, double, double]
Reducer 2
- Execution mode: llap
+ Execution mode: vectorized, llap
Reduce Vectorization:
enabled: true
enableConditionsMet: hive.vectorized.execution.reduce.enabled IS true, hive.execution.engine tez IN [tez, spark] IS true
- notVectorizedReason: GROUPBY operator: Vector aggregation : "var_pop" for input type: "STRUCT" and output type: "DOUBLE" and mode: FINAL not supported for evaluator GenericUDAFVarianceEvaluator
- vectorized: false
+ reduceColumnNullOrder:
+ reduceColumnSortOrder:
+ allNative: false
+ usesVectorUDFAdaptor: false
+ vectorized: true
+ rowBatchContext:
+ dataColumnCount: 10
+ dataColumns: VALUE._col0:double, VALUE._col1:double, VALUE._col2:bigint, VALUE._col3:double, VALUE._col4:tinyint, VALUE._col5:int, VALUE._col6:double, VALUE._col7:double, VALUE._col8:bigint, VALUE._col9:bigint
+ partitionColumnCount: 0
+ scratchColumnTypeNames: []
Reduce Operator Tree:
Group By Operator
- aggregations: var_pop(VALUE._col0), sum(VALUE._col1), max(VALUE._col2), max(VALUE._col3), var_samp(VALUE._col4), count(VALUE._col5)
+ aggregations: sum(VALUE._col0), sum(VALUE._col1), count(VALUE._col2), sum(VALUE._col3), max(VALUE._col4), max(VALUE._col5), sum(VALUE._col6), sum(VALUE._col7), count(VALUE._col8), count(VALUE._col9)
+ Group By Vectorization:
+ aggregators: VectorUDAFSumDouble(col 0:double) -> double, VectorUDAFSumDouble(col 1:double) -> double, VectorUDAFCountMerge(col 2:bigint) -> bigint, VectorUDAFSumDouble(col 3:double) -> double, VectorUDAFMaxLong(col 4:tinyint) -> tinyint, VectorUDAFMaxLong(col 5:int) -> int, VectorUDAFSumDouble(col 6:double) -> double, VectorUDAFSumDouble(col 7:double) -> double, VectorUDAFCountMerge(col 8:bigint) -> bigint, VectorUDAFCountMerge(col 9:bigint) -> bigint
+ className: VectorGroupByOperator
+ groupByMode: MERGEPARTIAL
+ native: false
+ vectorProcessingMode: GLOBAL
+ projectedOutputColumnNums: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
mode: mergepartial
- outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5
- Statistics: Num rows: 1 Data size: 40 Basic stats: COMPLETE Column stats: COMPLETE
+ outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8, _col9
+ Statistics: Num rows: 1 Data size: 72 Basic stats: COMPLETE Column stats: COMPLETE
Select Operator
- expressions: _col0 (type: double), (_col0 / -26.28D) (type: double), _col1 (type: double), (-1.389D + _col1) (type: double), (_col1 * (-1.389D + _col1)) (type: double), _col2 (type: tinyint), (- (_col1 * (-1.389D + _col1))) (type: double), _col3 (type: int), (CAST( _col3 AS decimal(10,0)) * 79.553) (type: decimal(16,3)), _col4 (type: double), (10.175D % (- (_col1 * (-1.389D + _col1)))) (type: double), _col5 (type: bigint), (-563 % _col3) (type: int)
+ expressions: ((_col0 - ((_col1 * _col1) / _col2)) / _col2) (type: double), (((_col0 - ((_col1 * _col1) / _col2)) / _col2) / -26.28D) (type: double), _col3 (type: double), (-1.389D + _col3) (type: double), (_col3 * (-1.389D + _col3)) (type: double), _col4 (type: tinyint), (- (_col3 * (-1.389D + _col3))) (type: double), _col5 (type: int), (CAST( _col5 AS decimal(10,0)) * 79.553) (type: decimal(16,3)), ((_col6 - ((_col7 * _col7) / _col8)) / CASE WHEN ((_col8 = 1L)) THEN (null) ELSE ((_col8 - 1)) END) (type: double), (10.175D % (- (_col3 * (-1.389D + _col3)))) (type: double), _col9 (type: bigint), (-563 % _col5) (type: int)
outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8, _col9, _col10, _col11, _col12
+ Select Vectorization:
+ className: VectorSelectOperator
+ native: true
+ projectedOutputColumnNums: [11, 10, 3, 12, 14, 4, 13, 5, 17, 18, 22, 9, 21]
+ selectExpressions: DoubleColDivideLongColumn(col 10:double, col 2:bigint)(children: DoubleColSubtractDoubleColumn(col 0:double, col 11:double)(children: DoubleColDivideLongColumn(col 10:double, col 2:bigint)(children: DoubleColMultiplyDoubleColumn(col 1:double, col 1:double) -> 10:double) -> 11:double) -> 10:double) -> 11:double, DoubleColDivideDoubleScalar(col 12:double, val -26.28)(children: DoubleColDivideLongColumn(col 10:double, col 2:bigint)(children: DoubleColSubtractDoubleColumn(col 0:double, col 12:double)(children: DoubleColDivideLongColumn(col 10:double, col 2:bigint)(children: DoubleColMultiplyDoubleColumn(col 1:double, col 1:double) -> 10:double) -> 12:double) -> 10:double) -> 12:double) -> 10:double, DoubleScalarAddDoubleColumn(val -1.389, col 3:double) -> 12:double, DoubleColMultiplyDoubleColumn(col 3:double, col 13:double)(children: DoubleScalarAddDoubleColumn(val -1.389, col 3:double) -> 13:double) -> 14:double, DoubleColUnaryMinus(col 15:doubl
e)(children: DoubleColMultiplyDoubleColumn(col 3:double, col 13:double)(children: DoubleScalarAddDoubleColumn(val -1.389, col 3:double) -> 13:double) -> 15:double) -> 13:double, DecimalColMultiplyDecimalScalar(col 16:decimal(10,0), val 79.553)(children: CastLongToDecimal(col 5:int) -> 16:decimal(10,0)) -> 17:decimal(16,3), DoubleColDivideLongColumn(col 15:double, col 21:bigint)(children: DoubleColSubtractDoubleColumn(col 6:double, col 18:double)(children: DoubleColDivideLongColumn(col 15:double, col 8:bigint)(children: DoubleColMultiplyDoubleColumn(col 7:double, col 7:double) -> 15:double) -> 18:double) -> 15:double, IfExprNullCondExpr(col 19:boolean, null, col 20:bigint)(children: LongColEqualLongScalar(col 8:bigint, val 1) -> 19:boolean, LongColSubtractLongScalar(col 8:bigint, val 1) -> 20:bigint) -> 21:bigint) -> 18:double, DoubleScalarModuloDoubleColumn(val 10.175, col 15:double)(children: DoubleColUnaryMinus(col 22:double)(children: DoubleColMultiplyDoubleColumn(col 3:double, c
ol 15:double)(children: DoubleScalarAddDoubleColumn(val -1.389, col 3:double) -> 15:double) -> 22:double) -> 15:double) -> 22:double, LongScalarModuloLongColumn(val -563, col 5:int) -> 21:int
Statistics: Num rows: 1 Data size: 196 Basic stats: COMPLETE Column stats: COMPLETE
File Output Operator
compressed: false
+ File Sink Vectorization:
+ className: VectorFileSinkOperator
+ native: false
Statistics: Num rows: 1 Data size: 196 Basic stats: COMPLETE Column stats: COMPLETE
table:
input format: org.apache.hadoop.mapred.SequenceFileInputFormat
@@ -193,4 +217,4 @@ WHERE (((cdouble > ctinyint)
POSTHOOK: type: QUERY
POSTHOOK: Input: default@alltypesorc
#### A masked pattern was here ####
-1074.830257547229 -40.89917266161449 -21997.674998402596 -21999.063998402595 4.839282601059194E8 62 -4.839282601059194E8 1073680599 85414512692.247 7.569848642620903E10 10.175 3745 -563
+1074.8302575472321 -40.899172661614614 -21997.674998402596 -21999.063998402595 4.839282601059194E8 62 -4.839282601059194E8 1073680599 85414512692.247 7.569848642620917E10 10.175 3745 -563
http://git-wip-us.apache.org/repos/asf/hive/blob/5cb8867b/ql/src/test/results/clientpositive/llap/vectorization_12.q.out
----------------------------------------------------------------------
diff --git a/ql/src/test/results/clientpositive/llap/vectorization_12.q.out b/ql/src/test/results/clientpositive/llap/vectorization_12.q.out
index c9faf55..0ead6c4 100644
--- a/ql/src/test/results/clientpositive/llap/vectorization_12.q.out
+++ b/ql/src/test/results/clientpositive/llap/vectorization_12.q.out
@@ -95,27 +95,28 @@ STAGE PLANS:
predicate: (((cboolean1 >= cboolean2) or (UDFToShort(ctinyint) <> csmallint)) and ((cstring1 like '%a') or ((cboolean2 <= 1) and (cbigint >= UDFToLong(csmallint)))) and ctimestamp1 is null) (type: boolean)
Statistics: Num rows: 1 Data size: 166 Basic stats: COMPLETE Column stats: COMPLETE
Select Operator
- expressions: cbigint (type: bigint), cdouble (type: double), cstring1 (type: string), cboolean1 (type: boolean)
- outputColumnNames: cbigint, cdouble, cstring1, cboolean1
+ expressions: cbigint (type: bigint), cboolean1 (type: boolean), cstring1 (type: string), cdouble (type: double), UDFToDouble(cbigint) (type: double), (UDFToDouble(cbigint) * UDFToDouble(cbigint)) (type: double), (cdouble * cdouble) (type: double)
+ outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6
Select Vectorization:
className: VectorSelectOperator
native: true
- projectedOutputColumnNums: [3, 5, 6, 10]
+ projectedOutputColumnNums: [3, 10, 6, 5, 13, 16, 14]
+ selectExpressions: CastLongToDouble(col 3:bigint) -> 13:double, DoubleColMultiplyDoubleColumn(col 14:double, col 15:double)(children: CastLongToDouble(col 3:bigint) -> 14:double, CastLongToDouble(col 3:bigint) -> 15:double) -> 16:double, DoubleColMultiplyDoubleColumn(col 5:double, col 5:double) -> 14:double
Statistics: Num rows: 1 Data size: 166 Basic stats: COMPLETE Column stats: COMPLETE
Group By Operator
- aggregations: count(cbigint), stddev_samp(cbigint), avg(cdouble), sum(cbigint), stddev_pop(cdouble)
+ aggregations: count(_col0), sum(_col5), sum(_col4), sum(_col3), count(_col3), sum(_col0), sum(_col6)
Group By Vectorization:
- aggregators: VectorUDAFCount(col 3:bigint) -> bigint, VectorUDAFVarLong(col 3:bigint) -> struct<count:bigint,sum:double,variance:double> aggregation: stddev_samp, VectorUDAFAvgDouble(col 5:double) -> struct<count:bigint,sum:double,input:double>, VectorUDAFSumLong(col 3:bigint) -> bigint, VectorUDAFVarDouble(col 5:double) -> struct<count:bigint,sum:double,variance:double> aggregation: stddev_pop
+ aggregators: VectorUDAFCount(col 3:bigint) -> bigint, VectorUDAFSumDouble(col 16:double) -> double, VectorUDAFSumDouble(col 13:double) -> double, VectorUDAFSumDouble(col 5:double) -> double, VectorUDAFCount(col 5:double) -> bigint, VectorUDAFSumLong(col 3:bigint) -> bigint, VectorUDAFSumDouble(col 14:double) -> double
className: VectorGroupByOperator
groupByMode: HASH
keyExpressions: col 5:double, col 3:bigint, col 6:string, col 10:boolean
native: false
vectorProcessingMode: HASH
- projectedOutputColumnNums: [0, 1, 2, 3, 4]
- keys: cdouble (type: double), cbigint (type: bigint), cstring1 (type: string), cboolean1 (type: boolean)
+ projectedOutputColumnNums: [0, 1, 2, 3, 4, 5, 6]
+ keys: _col3 (type: double), _col0 (type: bigint), _col2 (type: string), _col1 (type: boolean)
mode: hash
- outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8
- Statistics: Num rows: 1 Data size: 370 Basic stats: COMPLETE Column stats: COMPLETE
+ outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8, _col9, _col10
+ Statistics: Num rows: 1 Data size: 170 Basic stats: COMPLETE Column stats: COMPLETE
Reduce Output Operator
key expressions: _col0 (type: double), _col1 (type: bigint), _col2 (type: string), _col3 (type: boolean)
sort order: ++++
@@ -125,9 +126,9 @@ STAGE PLANS:
keyColumnNums: [0, 1, 2, 3]
native: true
nativeConditionsMet: hive.vectorized.execution.reducesink.new.enabled IS true, hive.execution.engine tez IN [tez, spark] IS true, No PTF TopN IS true, No DISTINCT columns IS true, BinarySortableSerDe for keys IS true, LazyBinarySerDe for values IS true
- valueColumnNums: [4, 5, 6, 7, 8]
- Statistics: Num rows: 1 Data size: 370 Basic stats: COMPLETE Column stats: COMPLETE
- value expressions: _col4 (type: bigint), _col5 (type: struct<count:bigint,sum:double,variance:double>), _col6 (type: struct<count:bigint,sum:double,input:double>), _col7 (type: bigint), _col8 (type: struct<count:bigint,sum:double,variance:double>)
+ valueColumnNums: [4, 5, 6, 7, 8, 9, 10]
+ Statistics: Num rows: 1 Data size: 170 Basic stats: COMPLETE Column stats: COMPLETE
+ value expressions: _col4 (type: bigint), _col5 (type: double), _col6 (type: double), _col7 (type: double), _col8 (type: bigint), _col9 (type: bigint), _col10 (type: double)
Execution mode: vectorized, llap
LLAP IO: all inputs
Map Vectorization:
@@ -144,29 +145,56 @@ STAGE PLANS:
includeColumns: [0, 1, 3, 5, 6, 8, 10, 11]
dataColumns: ctinyint:tinyint, csmallint:smallint, cint:int, cbigint:bigint, cfloat:float, cdouble:double, cstring1:string, cstring2:string, ctimestamp1:timestamp, ctimestamp2:timestamp, cboolean1:boolean, cboolean2:boolean
partitionColumnCount: 0
- scratchColumnTypeNames: []
+ scratchColumnTypeNames: [double, double, double, double]
Reducer 2
- Execution mode: llap
+ Execution mode: vectorized, llap
Reduce Vectorization:
enabled: true
enableConditionsMet: hive.vectorized.execution.reduce.enabled IS true, hive.execution.engine tez IN [tez, spark] IS true
- notVectorizedReason: GROUPBY operator: Vector aggregation : "stddev_samp" for input type: "STRUCT" and output type: "DOUBLE" and mode: FINAL not supported for evaluator GenericUDAFStdSampleEvaluator
- vectorized: false
+ reduceColumnNullOrder: aaaa
+ reduceColumnSortOrder: ++++
+ allNative: false
+ usesVectorUDFAdaptor: false
+ vectorized: true
+ rowBatchContext:
+ dataColumnCount: 11
+ dataColumns: KEY._col0:double, KEY._col1:bigint, KEY._col2:string, KEY._col3:boolean, VALUE._col0:bigint, VALUE._col1:double, VALUE._col2:double, VALUE._col3:double, VALUE._col4:bigint, VALUE._col5:bigint, VALUE._col6:double
+ partitionColumnCount: 0
+ scratchColumnTypeNames: []
Reduce Operator Tree:
Group By Operator
- aggregations: count(VALUE._col0), stddev_samp(VALUE._col1), avg(VALUE._col2), sum(VALUE._col3), stddev_pop(VALUE._col4)
+ aggregations: count(VALUE._col0), sum(VALUE._col1), sum(VALUE._col2), sum(VALUE._col3), count(VALUE._col4), sum(VALUE._col5), sum(VALUE._col6)
+ Group By Vectorization:
+ aggregators: VectorUDAFCountMerge(col 4:bigint) -> bigint, VectorUDAFSumDouble(col 5:double) -> double, VectorUDAFSumDouble(col 6:double) -> double, VectorUDAFSumDouble(col 7:double) -> double, VectorUDAFCountMerge(col 8:bigint) -> bigint, VectorUDAFSumLong(col 9:bigint) -> bigint, VectorUDAFSumDouble(col 10:double) -> double
+ className: VectorGroupByOperator
+ groupByMode: MERGEPARTIAL
+ keyExpressions: col 0:double, col 1:bigint, col 2:string, col 3:boolean
+ native: false
+ vectorProcessingMode: MERGE_PARTIAL
+ projectedOutputColumnNums: [0, 1, 2, 3, 4, 5, 6]
keys: KEY._col0 (type: double), KEY._col1 (type: bigint), KEY._col2 (type: string), KEY._col3 (type: boolean)
mode: mergepartial
- outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8
- Statistics: Num rows: 1 Data size: 154 Basic stats: COMPLETE Column stats: COMPLETE
+ outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8, _col9, _col10
+ Statistics: Num rows: 1 Data size: 170 Basic stats: COMPLETE Column stats: COMPLETE
Select Operator
- expressions: _col1 (type: bigint), _col3 (type: boolean), _col2 (type: string), _col0 (type: double), (-6432.0D * _col0) (type: double), (- _col1) (type: bigint), _col4 (type: bigint), (_col1 * _col4) (type: bigint), _col5 (type: double), ((-6432.0D * _col0) / -6432.0D) (type: double), (- ((-6432.0D * _col0) / -6432.0D)) (type: double), _col6 (type: double), (- (-6432.0D * _col0)) (type: double), (-5638.15 + CAST( _col1 AS decimal(19,0))) (type: decimal(22,2)), _col7 (type: bigint), (_col6 / (-6432.0D * _col0)) (type: double), (- (- ((-6432.0D * _col0) / -6432.0D))) (type: double), (((-6432.0D * _col0) / -6432.0D) + (- (-6432.0D * _col0))) (type: double), _col8 (type: double)
+ expressions: _col1 (type: bigint), _col3 (type: boolean), _col2 (type: string), _col0 (type: double), (-6432.0D * _col0) (type: double), (- _col1) (type: bigint), _col4 (type: bigint), (_col1 * _col4) (type: bigint), power(((_col5 - ((_col6 * _col6) / _col4)) / CASE WHEN ((_col4 = 1L)) THEN (null) ELSE ((_col4 - 1)) END), 0.5) (type: double), ((-6432.0D * _col0) / -6432.0D) (type: double), (- ((-6432.0D * _col0) / -6432.0D)) (type: double), (_col7 / _col8) (type: double), (- (-6432.0D * _col0)) (type: double), (-5638.15 + CAST( _col1 AS decimal(19,0))) (type: decimal(22,2)), _col9 (type: bigint), ((_col7 / _col8) / (-6432.0D * _col0)) (type: double), (- (- ((-6432.0D * _col0) / -6432.0D))) (type: double), (((-6432.0D * _col0) / -6432.0D) + (- (-6432.0D * _col0))) (type: double), power(((_col10 - ((_col7 * _col7) / _col8)) / _col8), 0.5) (type: double)
outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8, _col9, _col10, _col11, _col12, _col13, _col14, _col15, _col17, _col18, _col19
- Statistics: Num rows: 1 Data size: 338 Basic stats: COMPLETE Column stats: COMPLETE
+ Select Vectorization:
+ className: VectorSelectOperator
+ native: true
+ projectedOutputColumnNums: [1, 3, 2, 0, 11, 12, 4, 13, 14, 19, 15, 20, 22, 24, 9, 26, 25, 21, 27]
+ selectExpressions: DoubleScalarMultiplyDoubleColumn(val -6432.0, col 0:double) -> 11:double, LongColUnaryMinus(col 1:bigint) -> 12:bigint, LongColMultiplyLongColumn(col 1:bigint, col 4:bigint) -> 13:bigint, FuncPowerDoubleToDouble(col 15:double)(children: DoubleColDivideLongColumn(col 14:double, col 18:bigint)(children: DoubleColSubtractDoubleColumn(col 5:double, col 15:double)(children: DoubleColDivideLongColumn(col 14:double, col 4:bigint)(children: DoubleColMultiplyDoubleColumn(col 6:double, col 6:double) -> 14:double) -> 15:double) -> 14:double, IfExprNullCondExpr(col 16:boolean, null, col 17:bigint)(children: LongColEqualLongScalar(col 4:bigint, val 1) -> 16:boolean, LongColSubtractLongScalar(col 4:bigint, val 1) -> 17:bigint) -> 18:bigint) -> 15:double) -> 14:double, DoubleColDivideDoubleScalar(col 15:double, val -6432.0)(children: DoubleScalarMultiplyDoubleColumn(val -6432.0, col 0:double) -> 15:double) -> 19:double, DoubleColUnaryMinus(col 20:double)(ch
ildren: DoubleColDivideDoubleScalar(col 15:double, val -6432.0)(children: DoubleScalarMultiplyDoubleColumn(val -6432.0, col 0:double) -> 15:double) -> 20:double) -> 15:double, DoubleColDivideLongColumn(col 7:double, col 8:bigint) -> 20:double, DoubleColUnaryMinus(col 21:double)(children: DoubleScalarMultiplyDoubleColumn(val -6432.0, col 0:double) -> 21:double) -> 22:double, DecimalScalarAddDecimalColumn(val -5638.15, col 23:decimal(19,0))(children: CastLongToDecimal(col 1:bigint) -> 23:decimal(19,0)) -> 24:decimal(22,2), DoubleColDivideDoubleColumn(col 21:double, col 25:double)(children: DoubleColDivideLongColumn(col 7:double, col 8:bigint) -> 21:double, DoubleScalarMultiplyDoubleColumn(val -6432.0, col 0:double) -> 25:double) -> 26:double, DoubleColUnaryMinus(col 21:double)(children: DoubleColUnaryMinus(col 25:double)(children: DoubleColDivideDoubleScalar(col 21:double, val -6432.0)(children: DoubleScalarMultiplyDoubleColumn(val -6432.0, col 0:double) -> 21:double) -> 25:double) ->
21:double) -> 25:double, DoubleColAddDoubleColumn(col 27:double, col 28:double)(children: DoubleColDivideDoubleScalar(col 21:double, val -6432.0)(children: DoubleScalarMultiplyDoubleColumn(val -6432.0, col 0:double) -> 21:double) -> 27:double, DoubleColUnaryMinus(col 21:double)(children: DoubleScalarMultiplyDoubleColumn(val -6432.0, col 0:double) -> 21:double) -> 28:double) -> 21:double, FuncPowerDoubleToDouble(col 28:double)(children: DoubleColDivideLongColumn(col 27:double, col 8:bigint)(children: DoubleColSubtractDoubleColumn(col 10:double, col 28:double)(children: DoubleColDivideLongColumn(col 27:double, col 8:bigint)(children: DoubleColMultiplyDoubleColumn(col 7:double, col 7:double) -> 27:double) -> 28:double) -> 27:double) -> 28:double) -> 27:double
+ Statistics: Num rows: 1 Data size: 346 Basic stats: COMPLETE Column stats: COMPLETE
Reduce Output Operator
key expressions: _col3 (type: double), _col0 (type: bigint), _col2 (type: string)
sort order: +++
- Statistics: Num rows: 1 Data size: 338 Basic stats: COMPLETE Column stats: COMPLETE
+ Reduce Sink Vectorization:
+ className: VectorReduceSinkObjectHashOperator
+ keyColumnNums: [0, 1, 2]
+ native: true
+ nativeConditionsMet: hive.vectorized.execution.reducesink.new.enabled IS true, hive.execution.engine tez IN [tez, spark] IS true, No PTF TopN IS true, No DISTINCT columns IS true, BinarySortableSerDe for keys IS true, LazyBinarySerDe for values IS true
+ valueColumnNums: [3, 11, 12, 4, 13, 14, 19, 15, 20, 22, 24, 9, 26, 25, 21, 27]
+ Statistics: Num rows: 1 Data size: 346 Basic stats: COMPLETE Column stats: COMPLETE
value expressions: _col1 (type: boolean), _col4 (type: double), _col5 (type: bigint), _col6 (type: bigint), _col7 (type: bigint), _col8 (type: double), _col9 (type: double), _col10 (type: double), _col11 (type: double), _col12 (type: double), _col13 (type: decimal(22,2)), _col14 (type: bigint), _col15 (type: double), _col17 (type: double), _col18 (type: double), _col19 (type: double)
Reducer 3
Execution mode: vectorized, llap