diff --git a/ql/src/java/org/apache/hadoop/hive/ql/stats/StatsUtils.java b/ql/src/java/org/apache/hadoop/hive/ql/stats/StatsUtils.java index 42570ce..a069394 100644 --- a/ql/src/java/org/apache/hadoop/hive/ql/stats/StatsUtils.java +++ b/ql/src/java/org/apache/hadoop/hive/ql/stats/StatsUtils.java @@ -249,8 +249,14 @@ public static Statistics collectStatistics(HiveConf conf, PrunedPartitionList pa partNames.add(part.getName()); } neededColumns = processNeededColumns(schema, neededColumns); - AggrStats aggrStats = Hive.get().getAggrColStatsFor(table.getDbName(), table.getTableName(), - neededColumns, partNames); + AggrStats aggrStats = null; + // We check the sizes of neededColumns and partNames here. If either + // size is 0, aggrStats is null after several retries. Thus, we can + // skip the step to connect to the metastore. + if (neededColumns.size() > 0 && partNames.size() > 0) { + aggrStats = Hive.get().getAggrColStatsFor(table.getDbName(), table.getTableName(), + neededColumns, partNames); + } if (null == aggrStats || null == aggrStats.getColStats() || aggrStats.getColStatsSize() == 0) { // There are some partitions with no state (or we didn't fetch any state). diff --git a/ql/src/test/queries/clientpositive/stats_ppr_all.q b/ql/src/test/queries/clientpositive/stats_ppr_all.q new file mode 100644 index 0000000..b611e83 --- /dev/null +++ b/ql/src/test/queries/clientpositive/stats_ppr_all.q @@ -0,0 +1,24 @@ +set hive.stats.fetch.column.stats=true; + +drop table ss; + +CREATE TABLE ss ( + sales_order_id BIGINT, + order_amount FLOAT) +PARTITIONED BY (country STRING, year INT, month INT, day INT) stored as orc; + +insert into ss partition(country="US", year=2015, month=1, day=1) values(1,22.0); +insert into ss partition(country="US", year=2015, month=2, day=1) values(2,2.0); +insert into ss partition(country="US", year=2015, month=1, day=2) values(1,2.0); + +ANALYZE TABLE ss PARTITION(country,year,month,day) compute statistics for columns; + +explain select sum(order_amount) from ss where (country="US" and year=2015 and month=2 and day=1); + +explain select sum(order_amount) from ss where (year*10000+month*100+day) = "2015010" and 1>0; + +explain select sum(order_amount) from ss where (year*100+month*10+day) = "201511" and 1>0; + +explain select sum(order_amount) from ss where (year*100+month*10+day) > "201511" and 1>0; + +explain select '1' from ss where (year*100+month*10+day) > "201511"; \ No newline at end of file diff --git a/ql/src/test/results/clientpositive/stats_ppr_all.q.out b/ql/src/test/results/clientpositive/stats_ppr_all.q.out new file mode 100644 index 0000000..5f6f5d4 --- /dev/null +++ b/ql/src/test/results/clientpositive/stats_ppr_all.q.out @@ -0,0 +1,284 @@ +PREHOOK: query: drop table ss +PREHOOK: type: DROPTABLE +POSTHOOK: query: drop table ss +POSTHOOK: type: DROPTABLE +PREHOOK: query: CREATE TABLE ss ( + sales_order_id BIGINT, + order_amount FLOAT) +PARTITIONED BY (country STRING, year INT, month INT, day INT) stored as orc +PREHOOK: type: CREATETABLE +PREHOOK: Output: database:default +PREHOOK: Output: default@ss +POSTHOOK: query: CREATE TABLE ss ( + sales_order_id BIGINT, + order_amount FLOAT) +PARTITIONED BY (country STRING, year INT, month INT, day INT) stored as orc +POSTHOOK: type: CREATETABLE +POSTHOOK: Output: database:default +POSTHOOK: Output: default@ss +PREHOOK: query: insert into ss partition(country="US", year=2015, month=1, day=1) values(1,22.0) +PREHOOK: type: QUERY +PREHOOK: Input: default@values__tmp__table__1 +PREHOOK: Output: default@ss@country=US/year=2015/month=1/day=1 +POSTHOOK: query: insert into ss partition(country="US", year=2015, month=1, day=1) values(1,22.0) +POSTHOOK: type: QUERY +POSTHOOK: Input: default@values__tmp__table__1 +POSTHOOK: Output: default@ss@country=US/year=2015/month=1/day=1 +POSTHOOK: Lineage: ss PARTITION(country=US,year=2015,month=1,day=1).order_amount EXPRESSION [(values__tmp__table__1)values__tmp__table__1.FieldSchema(name:tmp_values_col2, type:string, comment:), ] +POSTHOOK: Lineage: ss PARTITION(country=US,year=2015,month=1,day=1).sales_order_id EXPRESSION [(values__tmp__table__1)values__tmp__table__1.FieldSchema(name:tmp_values_col1, type:string, comment:), ] +PREHOOK: query: insert into ss partition(country="US", year=2015, month=2, day=1) values(2,2.0) +PREHOOK: type: QUERY +PREHOOK: Input: default@values__tmp__table__2 +PREHOOK: Output: default@ss@country=US/year=2015/month=2/day=1 +POSTHOOK: query: insert into ss partition(country="US", year=2015, month=2, day=1) values(2,2.0) +POSTHOOK: type: QUERY +POSTHOOK: Input: default@values__tmp__table__2 +POSTHOOK: Output: default@ss@country=US/year=2015/month=2/day=1 +POSTHOOK: Lineage: ss PARTITION(country=US,year=2015,month=2,day=1).order_amount EXPRESSION [(values__tmp__table__2)values__tmp__table__2.FieldSchema(name:tmp_values_col2, type:string, comment:), ] +POSTHOOK: Lineage: ss PARTITION(country=US,year=2015,month=2,day=1).sales_order_id EXPRESSION [(values__tmp__table__2)values__tmp__table__2.FieldSchema(name:tmp_values_col1, type:string, comment:), ] +PREHOOK: query: insert into ss partition(country="US", year=2015, month=1, day=2) values(1,2.0) +PREHOOK: type: QUERY +PREHOOK: Input: default@values__tmp__table__3 +PREHOOK: Output: default@ss@country=US/year=2015/month=1/day=2 +POSTHOOK: query: insert into ss partition(country="US", year=2015, month=1, day=2) values(1,2.0) +POSTHOOK: type: QUERY +POSTHOOK: Input: default@values__tmp__table__3 +POSTHOOK: Output: default@ss@country=US/year=2015/month=1/day=2 +POSTHOOK: Lineage: ss PARTITION(country=US,year=2015,month=1,day=2).order_amount EXPRESSION [(values__tmp__table__3)values__tmp__table__3.FieldSchema(name:tmp_values_col2, type:string, comment:), ] +POSTHOOK: Lineage: ss PARTITION(country=US,year=2015,month=1,day=2).sales_order_id EXPRESSION [(values__tmp__table__3)values__tmp__table__3.FieldSchema(name:tmp_values_col1, type:string, comment:), ] +PREHOOK: query: ANALYZE TABLE ss PARTITION(country,year,month,day) compute statistics for columns +PREHOOK: type: QUERY +PREHOOK: Input: default@ss +PREHOOK: Input: default@ss@country=US/year=2015/month=1/day=1 +PREHOOK: Input: default@ss@country=US/year=2015/month=1/day=2 +PREHOOK: Input: default@ss@country=US/year=2015/month=2/day=1 +#### A masked pattern was here #### +POSTHOOK: query: ANALYZE TABLE ss PARTITION(country,year,month,day) compute statistics for columns +POSTHOOK: type: QUERY +POSTHOOK: Input: default@ss +POSTHOOK: Input: default@ss@country=US/year=2015/month=1/day=1 +POSTHOOK: Input: default@ss@country=US/year=2015/month=1/day=2 +POSTHOOK: Input: default@ss@country=US/year=2015/month=2/day=1 +#### A masked pattern was here #### +PREHOOK: query: explain select sum(order_amount) from ss where (country="US" and year=2015 and month=2 and day=1) +PREHOOK: type: QUERY +POSTHOOK: query: explain select sum(order_amount) from ss where (country="US" and year=2015 and month=2 and day=1) +POSTHOOK: type: QUERY +STAGE DEPENDENCIES: + Stage-1 is a root stage + Stage-0 depends on stages: Stage-1 + +STAGE PLANS: + Stage: Stage-1 + Map Reduce + Map Operator Tree: + TableScan + alias: ss + Statistics: Num rows: 1 Data size: 12 Basic stats: COMPLETE Column stats: COMPLETE + Select Operator + expressions: order_amount (type: float) + outputColumnNames: _col0 + Statistics: Num rows: 1 Data size: 12 Basic stats: COMPLETE Column stats: COMPLETE + Group By Operator + aggregations: sum(_col0) + mode: hash + outputColumnNames: _col0 + Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: COMPLETE + Reduce Output Operator + sort order: + Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: COMPLETE + value expressions: _col0 (type: double) + Reduce Operator Tree: + Group By Operator + aggregations: sum(VALUE._col0) + mode: mergepartial + outputColumnNames: _col0 + Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: COMPLETE + File Output Operator + compressed: false + Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: COMPLETE + table: + input format: org.apache.hadoop.mapred.TextInputFormat + output format: org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat + serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe + + Stage: Stage-0 + Fetch Operator + limit: -1 + Processor Tree: + ListSink + +PREHOOK: query: explain select sum(order_amount) from ss where (year*10000+month*100+day) = "2015010" and 1>0 +PREHOOK: type: QUERY +POSTHOOK: query: explain select sum(order_amount) from ss where (year*10000+month*100+day) = "2015010" and 1>0 +POSTHOOK: type: QUERY +STAGE DEPENDENCIES: + Stage-1 is a root stage + Stage-0 depends on stages: Stage-1 + +STAGE PLANS: + Stage: Stage-1 + Map Reduce + Map Operator Tree: + TableScan + alias: ss + Statistics: Num rows: 1 Data size: 0 Basic stats: PARTIAL Column stats: PARTIAL + Filter Operator + predicate: (UDFToDouble((((year * 10000) + (month * 100)) + day)) = 2015010.0) (type: boolean) + Statistics: Num rows: 1 Data size: 0 Basic stats: PARTIAL Column stats: PARTIAL + Select Operator + expressions: order_amount (type: float) + outputColumnNames: _col0 + Statistics: Num rows: 1 Data size: 0 Basic stats: PARTIAL Column stats: PARTIAL + Group By Operator + aggregations: sum(_col0) + mode: hash + outputColumnNames: _col0 + Statistics: Num rows: 1 Data size: 20 Basic stats: COMPLETE Column stats: PARTIAL + Reduce Output Operator + sort order: + Statistics: Num rows: 1 Data size: 20 Basic stats: COMPLETE Column stats: PARTIAL + value expressions: _col0 (type: double) + Reduce Operator Tree: + Group By Operator + aggregations: sum(VALUE._col0) + mode: mergepartial + outputColumnNames: _col0 + Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: PARTIAL + File Output Operator + compressed: false + Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: PARTIAL + table: + input format: org.apache.hadoop.mapred.TextInputFormat + output format: org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat + serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe + + Stage: Stage-0 + Fetch Operator + limit: -1 + Processor Tree: + ListSink + +PREHOOK: query: explain select sum(order_amount) from ss where (year*100+month*10+day) = "201511" and 1>0 +PREHOOK: type: QUERY +POSTHOOK: query: explain select sum(order_amount) from ss where (year*100+month*10+day) = "201511" and 1>0 +POSTHOOK: type: QUERY +STAGE DEPENDENCIES: + Stage-1 is a root stage + Stage-0 depends on stages: Stage-1 + +STAGE PLANS: + Stage: Stage-1 + Map Reduce + Map Operator Tree: + TableScan + alias: ss + Statistics: Num rows: 1 Data size: 12 Basic stats: COMPLETE Column stats: COMPLETE + Select Operator + expressions: order_amount (type: float) + outputColumnNames: _col0 + Statistics: Num rows: 1 Data size: 12 Basic stats: COMPLETE Column stats: COMPLETE + Group By Operator + aggregations: sum(_col0) + mode: hash + outputColumnNames: _col0 + Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: COMPLETE + Reduce Output Operator + sort order: + Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: COMPLETE + value expressions: _col0 (type: double) + Reduce Operator Tree: + Group By Operator + aggregations: sum(VALUE._col0) + mode: mergepartial + outputColumnNames: _col0 + Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: COMPLETE + File Output Operator + compressed: false + Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: COMPLETE + table: + input format: org.apache.hadoop.mapred.TextInputFormat + output format: org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat + serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe + + Stage: Stage-0 + Fetch Operator + limit: -1 + Processor Tree: + ListSink + +PREHOOK: query: explain select sum(order_amount) from ss where (year*100+month*10+day) > "201511" and 1>0 +PREHOOK: type: QUERY +POSTHOOK: query: explain select sum(order_amount) from ss where (year*100+month*10+day) > "201511" and 1>0 +POSTHOOK: type: QUERY +STAGE DEPENDENCIES: + Stage-1 is a root stage + Stage-0 depends on stages: Stage-1 + +STAGE PLANS: + Stage: Stage-1 + Map Reduce + Map Operator Tree: + TableScan + alias: ss + Statistics: Num rows: 2 Data size: 24 Basic stats: COMPLETE Column stats: COMPLETE + Filter Operator + predicate: (UDFToDouble(((201500 + (month * 10)) + day)) > 201511.0) (type: boolean) + Statistics: Num rows: 1 Data size: 12 Basic stats: COMPLETE Column stats: COMPLETE + Select Operator + expressions: order_amount (type: float) + outputColumnNames: _col0 + Statistics: Num rows: 1 Data size: 12 Basic stats: COMPLETE Column stats: COMPLETE + Group By Operator + aggregations: sum(_col0) + mode: hash + outputColumnNames: _col0 + Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: COMPLETE + Reduce Output Operator + sort order: + Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: COMPLETE + value expressions: _col0 (type: double) + Reduce Operator Tree: + Group By Operator + aggregations: sum(VALUE._col0) + mode: mergepartial + outputColumnNames: _col0 + Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: COMPLETE + File Output Operator + compressed: false + Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: COMPLETE + table: + input format: org.apache.hadoop.mapred.TextInputFormat + output format: org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat + serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe + + Stage: Stage-0 + Fetch Operator + limit: -1 + Processor Tree: + ListSink + +PREHOOK: query: explain select '1' from ss where (year*100+month*10+day) > "201511" +PREHOOK: type: QUERY +POSTHOOK: query: explain select '1' from ss where (year*100+month*10+day) > "201511" +POSTHOOK: type: QUERY +STAGE DEPENDENCIES: + Stage-0 is a root stage + +STAGE PLANS: + Stage: Stage-0 + Fetch Operator + limit: -1 + Processor Tree: + TableScan + alias: ss + Statistics: Num rows: 2 Data size: 24 Basic stats: COMPLETE Column stats: COMPLETE + Filter Operator + predicate: (UDFToDouble(((201500 + (month * 10)) + day)) > 201511.0) (type: boolean) + Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: COMPLETE + Select Operator + expressions: '1' (type: string) + outputColumnNames: _col0 + Statistics: Num rows: 1 Data size: 85 Basic stats: COMPLETE Column stats: COMPLETE + ListSink +