jprdonnelly/impala
1
0
Fork 0
mirror of https://github.com/apache/impala.git synced 2026-01-04 09:00:56 -05:00
Code Issues Projects Releases Wiki Activity
Files
e94de02469a3dfeda7d9358d7a78cb54c7a67159
impala/testdata/workloads/functional-planner/queries/PlannerTest/union.test
Alex Behm 881f3a8c33 Re-order union operands descending by their estimated per-host memory. 
Re-order union operands descending by their estimated per-host memory,
s.t. parent nodes can gauge the peak memory consumption of a MergeNode after
opening it during execution (a MergeNode opens its first operand in Open()).
Scan nodes are always ordered last because they can dynamically scale down their
memory usage, whereas many other nodes cannot (e.g., joins, aggregations).
One goal is to decrease the likelihood of a SortNode parent claiming too much
memory in its Open(), possibly causing the mem limit to be hit when subsequent
union operands are executed.

Change-Id: Ia51caaffd55305ea3dbd2146cd55acc7da67f382
Reviewed-on: http://gerrit.ent.cloudera.com:8080/3146
Reviewed-by: Alex Behm <alex.behm@cloudera.com>
Tested-by: Alex Behm <alex.behm@cloudera.com>
Reviewed-on: http://gerrit.ent.cloudera.com:8080/3213
Tested-by: jenkins
2014-06-20 18:46:10 -07:00

2650 lines
82 KiB
Plaintext
Raw Blame History

# binding predicates of b.month equiv class gets propagated into union
select * from
(select year, month from functional.alltypes
union all
select year, month from functional.alltypes) a
inner join
functional.alltypessmall b
on (a.month = b.month)
where b.month = 1
---- PLAN
04:HASH JOIN [INNER JOIN]
| hash predicates: month = b.month
|
|--03:SCAN HDFS [functional.alltypessmall b]
| partitions=1/4 size=1.57KB compact
|
00:UNION
|
|--02:SCAN HDFS [functional.alltypes]
| partitions=2/24 size=40.32KB
|
01:SCAN HDFS [functional.alltypes]
partitions=2/24 size=40.32KB
---- DISTRIBUTEDPLAN
06:EXCHANGE [UNPARTITIONED]
|
04:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: month = b.month
|
|--05:EXCHANGE [BROADCAST]
| |
| 03:SCAN HDFS [functional.alltypessmall b]
| partitions=1/4 size=1.57KB
|
00:UNION
|
|--02:SCAN HDFS [functional.alltypes]
| partitions=2/24 size=40.32KB
|
01:SCAN HDFS [functional.alltypes]
partitions=2/24 size=40.32KB
====
// Only UNION ALL, no nested unions
select * from functional.alltypestiny where year=2009 and month=1
union all
select * from functional.alltypestiny where year=2009 and month=1
union all
select * from functional.alltypestiny where year=2009 and month=2
---- PLAN
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
04:EXCHANGE [UNPARTITIONED]
|
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Only UNION ALL with limit inside operands. One of the operands also has an order by.
select * from functional.alltypestiny where year=2009 and month=1 limit 1
union all
select * from functional.alltypestiny where year=2009 and month=1 order by int_col limit 1
union all
select * from functional.alltypestiny where year=2009 and month=2 limit 1
---- PLAN
00:UNION
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
| limit: 1
|
|--03:TOP-N [LIMIT=1]
| | order by: int_col ASC
| |
| 02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
limit: 1
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 4:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
00:UNION
|
|--07:EXCHANGE [UNPARTITIONED]
| | limit: 1
| |
| 04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
| limit: 1
|
|--06:MERGING-EXCHANGE [UNPARTITIONED]
| | order by: int_col ASC
| | limit: 1
| |
| 03:TOP-N [LIMIT=1]
| | order by: int_col ASC
| |
| 02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
05:EXCHANGE [UNPARTITIONED]
| limit: 1
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
limit: 1
====
// Only UNION DISTINCT, no nested unions
select * from functional.alltypestiny where year=2009 and month=1
union distinct
select * from functional.alltypestiny where year=2009 and month=1
union distinct
select * from functional.alltypestiny where year=2009 and month=2
---- PLAN
04:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
07:EXCHANGE [UNPARTITIONED]
|
06:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
05:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
04:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Only UNION ALL, mixed selects with and without from clauses, no nested unions
select * from functional.alltypestiny where year=2009 and month=1
union all
select 0,true,0,0,0,0,0,0,'01/01/09','0',cast('2009-01-01 00:00:00' as timestamp),2009,1
union all
select * from functional.alltypestiny where year=2009 and month=1
union all
select 1,false,1,1,1,10,1.1,10.1,'01/01/09','1',cast('2009-01-01 00:01:00' as timestamp),2009,1
---- PLAN
00:UNION
| constant-operands=2
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
---- DISTRIBUTEDPLAN
03:EXCHANGE [UNPARTITIONED]
|
00:UNION
| constant-operands=2
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Only UNION DISTINCT, mixed selects with and without from clauses, no nested unions
select * from functional.alltypestiny where year=2009 and month=1
union distinct
select 0,true,0,0,0,0,0,0,'01/01/09','0',cast('2009-01-01 00:00:00' as timestamp),2009,1
union distinct
select * from functional.alltypestiny where year=2009 and month=1
union distinct
select 1,false,1,1,1,10,1.1,10.1,'01/01/09','1',cast('2009-01-01 00:01:00' as timestamp),2009,1
---- PLAN
03:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
| constant-operands=2
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
---- DISTRIBUTEDPLAN
06:EXCHANGE [UNPARTITIONED]
|
05:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
04:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
03:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
| constant-operands=2
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Mixed UNION ALL/DISTINCT but effectively only UNION DISTINCT, no nested unions,
// with order by and limit
select * from functional.alltypestiny where year=2009 and month=1
union all
select * from functional.alltypestiny where year=2009 and month=1
union all
select * from functional.alltypestiny where year=2009 and month=2
union distinct
(select * from functional.alltypestiny where year=2009 and month=2)
order by 3 limit 3
---- PLAN
06:TOP-N [LIMIT=3]
| order by: tinyint_col ASC
|
05:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 4:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
09:MERGING-EXCHANGE [UNPARTITIONED]
| order by: tinyint_col ASC
| limit: 3
|
06:TOP-N [LIMIT=3]
| order by: tinyint_col ASC
|
08:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
07:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
05:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Mixed UNION ALL/DISTINCT, no nested unions, with order by and limit
select * from functional.alltypestiny where year=2009 and month=1
union distinct
select * from functional.alltypestiny where year=2009 and month=1
union all
select * from functional.alltypestiny where year=2009 and month=2
union all
(select * from functional.alltypestiny where year=2009 and month=2)
order by 3,4 limit 3
---- PLAN
07:TOP-N [LIMIT=3]
| order by: tinyint_col ASC, smallint_col ASC
|
04:UNION
|
|--03:AGGREGATE [FINALIZE]
| | group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
| |
| 00:UNION
| |
| |--02:SCAN HDFS [functional.alltypestiny]
| | partitions=1/4 size=115B
| |
| 01:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--06:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
05:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 5:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 6:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
10:MERGING-EXCHANGE [UNPARTITIONED]
| order by: tinyint_col ASC, smallint_col ASC
| limit: 3
|
07:TOP-N [LIMIT=3]
| order by: tinyint_col ASC, smallint_col ASC
|
04:UNION
|
|--06:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--05:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
09:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
08:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
03:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Mixed UNION ALL/DISTINCT, no nested unions, with order by and limit
select * from functional.alltypestiny where year=2009 and month=1
union all
select * from functional.alltypestiny where year=2009 and month=1
union distinct
select * from functional.alltypestiny where year=2009 and month=2
union all
(select * from functional.alltypestiny where year=2009 and month=2)
order by 3,4 limit 4
---- PLAN
07:TOP-N [LIMIT=4]
| order by: tinyint_col ASC, smallint_col ASC
|
05:UNION
|
|--04:AGGREGATE [FINALIZE]
| | group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
| |
| 00:UNION
| |
| |--03:SCAN HDFS [functional.alltypestiny]
| | partitions=1/4 size=115B
| |
| |--02:SCAN HDFS [functional.alltypestiny]
| | partitions=1/4 size=115B
| |
| 01:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
06:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 6:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
10:MERGING-EXCHANGE [UNPARTITIONED]
| order by: tinyint_col ASC, smallint_col ASC
| limit: 4
|
07:TOP-N [LIMIT=4]
| order by: tinyint_col ASC, smallint_col ASC
|
05:UNION
|
|--06:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
09:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
08:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
04:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: Only UNION ALL, first operand is nested
(select * from functional.alltypestiny where year=2009 and month=1
union all
select * from functional.alltypestiny where year=2009 and month=2)
union all
select * from functional.alltypestiny where year=2009 and month=1
---- PLAN
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
---- DISTRIBUTEDPLAN
04:EXCHANGE [UNPARTITIONED]
|
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: Only UNION ALL, second operand is nested
select * from functional.alltypestiny where year=2009 and month=1
union all
(select * from functional.alltypestiny where year=2009 and month=1
union all
select * from functional.alltypestiny where year=2009 and month=2)
---- PLAN
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
04:EXCHANGE [UNPARTITIONED]
|
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: Only UNION DISTINCT, first operand is nested
(select * from functional.alltypestiny where year=2009 and month=1
union distinct
select * from functional.alltypestiny where year=2009 and month=2)
union distinct
select * from functional.alltypestiny where year=2009 and month=1
---- PLAN
04:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
---- DISTRIBUTEDPLAN
07:EXCHANGE [UNPARTITIONED]
|
06:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
05:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
04:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: Only UNION DISTINCT, second operand is nested
select * from functional.alltypestiny where year=2009 and month=1
union distinct
(select * from functional.alltypestiny where year=2009 and month=1
union distinct
select * from functional.alltypestiny where year=2009 and month=2)
---- PLAN
04:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
07:EXCHANGE [UNPARTITIONED]
|
06:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
05:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
04:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: UNION ALL doesn't absorb nested union with DISTINCT,
// first operand is nested
(select * from functional.alltypestiny where year=2009 and month=1
union distinct
select * from functional.alltypestiny where year=2009 and month=2)
union all
select * from functional.alltypestiny where year=2009 and month=1
---- PLAN
00:UNION
|
|--05:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
04:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
01:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
02:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 5:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
---- DISTRIBUTEDPLAN
08:EXCHANGE [UNPARTITIONED]
|
00:UNION
|
|--05:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
07:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
06:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
04:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
01:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
02:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: UNION ALL doesn't absorb nested union with DISTINCT,
// second operand is nested
select * from functional.alltypestiny where year=2009 and month=1
union all
(select * from functional.alltypestiny where year=2009 and month=1
union distinct
select * from functional.alltypestiny where year=2009 and month=2)
---- PLAN
00:UNION
|
|--05:AGGREGATE [FINALIZE]
| | group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
| |
| 02:UNION
| |
| |--04:SCAN HDFS [functional.alltypestiny]
| | partitions=1/4 size=115B
| |
| 03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 4:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
08:EXCHANGE [UNPARTITIONED]
|
00:UNION
|
|--01:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
07:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
06:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
05:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
02:UNION
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
03:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: UNION ALL absorbs the children but not directly the operands
// of a nested union with mixed ALL/DISTINCT, first operand is nested
(select * from functional.alltypestiny where year=2009 and month=1
union distinct
select * from functional.alltypestiny where year=2009 and month=2
union all
select * from functional.alltypestiny where year=2009 and month=2)
union all
select * from functional.alltypestiny where year=2009 and month=1
---- PLAN
00:UNION
|
|--06:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--05:AGGREGATE [FINALIZE]
| | group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
| |
| 02:UNION
| |
| |--04:SCAN HDFS [functional.alltypestiny]
| | partitions=1/4 size=115B
| |
| 03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 4:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 6:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
---- DISTRIBUTEDPLAN
09:EXCHANGE [UNPARTITIONED]
|
00:UNION
|
|--06:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--01:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
08:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
07:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
05:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
02:UNION
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
03:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: UNION ALL absorbs the children but not directly the operands
// of a nested union with mixed ALL/DISTINCT, second operand is nested
select * from functional.alltypestiny where year=2009 and month=1
union all
(select * from functional.alltypestiny where year=2009 and month=1
union distinct
select * from functional.alltypestiny where year=2009 and month=2
union all
select * from functional.alltypestiny where year=2009 and month=2)
---- PLAN
00:UNION
|
|--06:AGGREGATE [FINALIZE]
| | group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
| |
| 03:UNION
| |
| |--05:SCAN HDFS [functional.alltypestiny]
| | partitions=1/4 size=115B
| |
| 04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 4:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 5:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
09:EXCHANGE [UNPARTITIONED]
|
00:UNION
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--01:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
08:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
07:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
06:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
03:UNION
|
|--05:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
04:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: UNION ALL doesn't absorb the children of a nested union
// with mixed ALL/DISTINCT and limit, second operand is nested
select * from functional.alltypestiny where year=2009 and month=1
union all
(select * from functional.alltypestiny where year=2009 and month=1
union distinct
select * from functional.alltypestiny where year=2009 and month=2
union all
(select * from functional.alltypestiny where year=2009 and month=2)
limit 10)
---- PLAN
00:UNION
|
|--06:UNION
| | limit: 10
| |
| |--05:AGGREGATE [FINALIZE]
| | | group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
| | |
| | 02:UNION
| | |
| | |--04:SCAN HDFS [functional.alltypestiny]
| | | partitions=1/4 size=115B
| | |
| | 03:SCAN HDFS [functional.alltypestiny]
| | partitions=1/4 size=115B
| |
| 07:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 4:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 7:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
12:EXCHANGE [UNPARTITIONED]
|
00:UNION
|
|--01:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
11:EXCHANGE [RANDOM]
|
10:EXCHANGE [UNPARTITIONED]
| limit: 10
|
06:UNION
| limit: 10
|
|--07:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
09:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
08:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
05:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
02:UNION
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
03:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: UNION ALL doesn't absorb nested union with order by and limit,
// first operand is nested
(select * from functional.alltypestiny where year=2009 and month=1
union all
(select * from functional.alltypestiny where year=2009 and month=2)
order by 3 limit 3)
union all
select * from functional.alltypestiny where year=2009 and month=1
---- PLAN
00:UNION
|
|--05:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
04:TOP-N [LIMIT=3]
| order by: tinyint_col ASC
|
01:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
02:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 5:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
---- DISTRIBUTEDPLAN
08:EXCHANGE [UNPARTITIONED]
|
00:UNION
|
|--05:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
07:EXCHANGE [RANDOM]
|
06:MERGING-EXCHANGE [UNPARTITIONED]
| order by: tinyint_col ASC
| limit: 3
|
04:TOP-N [LIMIT=3]
| order by: tinyint_col ASC
|
01:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
02:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: UNION ALL doesn't absorb nested union with order by and limit,
// second operand is nested
select * from functional.alltypestiny where year=2009 and month=1
union all
(select * from functional.alltypestiny where year=2009 and month=1
union all
(select * from functional.alltypestiny where year=2009 and month=2)
order by 3 limit 3)
---- PLAN
00:UNION
|
|--05:TOP-N [LIMIT=3]
| | order by: tinyint_col ASC
| |
| 02:UNION
| |
| |--04:SCAN HDFS [functional.alltypestiny]
| | partitions=1/4 size=115B
| |
| 03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 4:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
08:EXCHANGE [UNPARTITIONED]
|
00:UNION
|
|--01:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
07:EXCHANGE [RANDOM]
|
06:MERGING-EXCHANGE [UNPARTITIONED]
| order by: tinyint_col ASC
| limit: 3
|
05:TOP-N [LIMIT=3]
| order by: tinyint_col ASC
|
02:UNION
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
03:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: UNION DISTINCT absorbs nested union with ALL
// first operand is nested
(select * from functional.alltypestiny where year=2009 and month=1
union all
select * from functional.alltypestiny where year=2009 and month=2)
union distinct
select * from functional.alltypestiny where year=2009 and month=1
---- PLAN
04:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
---- DISTRIBUTEDPLAN
07:EXCHANGE [UNPARTITIONED]
|
06:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
05:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
04:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: UNION DISTINCT absorbs nested union with ALL,
// second operand is nested
select * from functional.alltypestiny where year=2009 and month=1
union distinct
(select * from functional.alltypestiny where year=2009 and month=1
union all
select * from functional.alltypestiny where year=2009 and month=2)
---- PLAN
04:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
07:EXCHANGE [UNPARTITIONED]
|
06:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
05:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
04:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: UNION DISTINCT absorbs nested union with mixed ALL/DISTINCT,
// first operand is nested
(select * from functional.alltypestiny where year=2009 and month=1
union distinct
select * from functional.alltypestiny where year=2009 and month=2
union all
select * from functional.alltypestiny where year=2009 and month=2)
union distinct
select * from functional.alltypestiny where year=2009 and month=1
---- PLAN
05:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 4:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
---- DISTRIBUTEDPLAN
08:EXCHANGE [UNPARTITIONED]
|
07:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
06:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
05:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: UNION DISTINCT absorbs nested union with mixed ALL/DISTINCT,
// second operand is nested
select * from functional.alltypestiny where year=2009 and month=1
union distinct
(select * from functional.alltypestiny where year=2009 and month=1
union distinct
select * from functional.alltypestiny where year=2009 and month=2
union all
select * from functional.alltypestiny where year=2009 and month=2)
---- PLAN
05:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 4:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
08:EXCHANGE [UNPARTITIONED]
|
07:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
06:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
05:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: UNION DISTINCT doesn't absorb nested union with order by and limit,
// first operand is nested
(select * from functional.alltypestiny where year=2009 and month=1
union all
(select * from functional.alltypestiny where year=2009 and month=2)
order by 3 limit 3)
union distinct
select * from functional.alltypestiny where year=2009 and month=1
---- PLAN
06:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--05:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
04:TOP-N [LIMIT=3]
| order by: tinyint_col ASC
|
01:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
02:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 5:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
---- DISTRIBUTEDPLAN
11:EXCHANGE [UNPARTITIONED]
|
10:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
09:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
06:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--05:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
08:EXCHANGE [RANDOM]
|
07:MERGING-EXCHANGE [UNPARTITIONED]
| order by: tinyint_col ASC
| limit: 3
|
04:TOP-N [LIMIT=3]
| order by: tinyint_col ASC
|
01:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
02:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Union unnesting: UNION DISTINCT doesn't absorb nested union with order by and limit
// second operand is nested
select * from functional.alltypestiny where year=2009 and month=1
union distinct
(select * from functional.alltypestiny where year=2009 and month=1
union all
(select * from functional.alltypestiny where year=2009 and month=2)
order by 3 limit 3)
---- PLAN
06:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--05:TOP-N [LIMIT=3]
| | order by: tinyint_col ASC
| |
| 02:UNION
| |
| |--04:SCAN HDFS [functional.alltypestiny]
| | partitions=1/4 size=115B
| |
| 03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 4:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
11:EXCHANGE [UNPARTITIONED]
|
10:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
09:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
06:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--01:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
08:EXCHANGE [RANDOM]
|
07:MERGING-EXCHANGE [UNPARTITIONED]
| order by: tinyint_col ASC
| limit: 3
|
05:TOP-N [LIMIT=3]
| order by: tinyint_col ASC
|
02:UNION
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
03:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Complex union unnesting: Multiple levels of UNION ALL, fully unnestable
select * from functional.alltypestiny where year=2009 and month=1
union all
(select * from functional.alltypestiny where year=2009 and month=1
union all
(select * from functional.alltypestiny where year=2009 and month=2
union all
(select * from functional.alltypestiny where year=2009 and month=2
union all
select * from functional.alltypestiny where year=2009 and month=3)))
---- PLAN
00:UNION
|
|--05:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 4:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 5:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=3/090301.txt 0:115
---- DISTRIBUTEDPLAN
06:EXCHANGE [UNPARTITIONED]
|
00:UNION
|
|--05:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Complex union unnesting: Multiple levels of UNION DISTINCT, fully unnestable
select * from functional.alltypestiny where year=2009 and month=1
union distinct
(select * from functional.alltypestiny where year=2009 and month=1
union distinct
(select * from functional.alltypestiny where year=2009 and month=2
union distinct
(select * from functional.alltypestiny where year=2009 and month=2
union distinct
select * from functional.alltypestiny where year=2009 and month=3)))
---- PLAN
06:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--05:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 4:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 5:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=3/090301.txt 0:115
---- DISTRIBUTEDPLAN
09:EXCHANGE [UNPARTITIONED]
|
08:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
07:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
06:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--05:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Complex union unnesting: Partially unnestable up to 2nd level
select * from functional.alltypestiny where year=2009 and month=1
union all
(select * from functional.alltypestiny where year=2009 and month=1
union distinct
(select * from functional.alltypestiny where year=2009 and month=2
union all
(select * from functional.alltypestiny where year=2009 and month=2
union distinct
(select * from functional.alltypestiny where year=2009 and month=3)
order by 3 limit 3)))
---- PLAN
00:UNION
|
|--10:AGGREGATE [FINALIZE]
| | group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
| |
| 02:UNION
| |
| |--09:TOP-N [LIMIT=3]
| | | order by: tinyint_col ASC
| | |
| | 08:AGGREGATE [FINALIZE]
| | | group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
| | |
| | 05:UNION
| | |
| | |--07:SCAN HDFS [functional.alltypestiny]
| | | partitions=1/4 size=115B
| | |
| | 06:SCAN HDFS [functional.alltypestiny]
| | partitions=1/4 size=115B
| |
| |--04:SCAN HDFS [functional.alltypestiny]
| | partitions=1/4 size=115B
| |
| 03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 4:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 6:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 7:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=3/090301.txt 0:115
---- DISTRIBUTEDPLAN
17:EXCHANGE [UNPARTITIONED]
|
00:UNION
|
|--01:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
16:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
15:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
10:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
02:UNION
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
14:EXCHANGE [RANDOM]
|
13:MERGING-EXCHANGE [UNPARTITIONED]
| order by: tinyint_col ASC
| limit: 3
|
09:TOP-N [LIMIT=3]
| order by: tinyint_col ASC
|
12:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
11:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
08:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
05:UNION
|
|--07:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
06:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Complex union unnesting: Partially unnestable up to 1st level
select * from functional.alltypestiny where year=2009 and month=1
union distinct
(select * from functional.alltypestiny where year=2009 and month=1
union distinct
(select * from functional.alltypestiny where year=2009 and month=2
union all
(select * from functional.alltypestiny where year=2009 and month=2
union distinct
(select * from functional.alltypestiny where year=2009 and month=3)
order by 3 limit 3)))
---- PLAN
09:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--08:TOP-N [LIMIT=3]
| | order by: tinyint_col ASC
| |
| 07:AGGREGATE [FINALIZE]
| | group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
| |
| 04:UNION
| |
| |--06:SCAN HDFS [functional.alltypestiny]
| | partitions=1/4 size=115B
| |
| 05:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 5:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 6:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=3/090301.txt 0:115
---- DISTRIBUTEDPLAN
16:EXCHANGE [UNPARTITIONED]
|
15:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
14:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
09:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--01:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
13:EXCHANGE [RANDOM]
|
12:MERGING-EXCHANGE [UNPARTITIONED]
| order by: tinyint_col ASC
| limit: 3
|
08:TOP-N [LIMIT=3]
| order by: tinyint_col ASC
|
11:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
10:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
07:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
04:UNION
|
|--06:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
05:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// Complex union unnesting: Multiple nested unions to test all rules in a single query
select * from functional.alltypestiny where year=2009 and month=1
union distinct
(select * from functional.alltypestiny where year=2009 and month=1
union all
select * from functional.alltypestiny where year=2009 and month=2)
union distinct
(select * from functional.alltypestiny where year=2009 and month=2
union all
(select * from functional.alltypestiny where year=2009 and month=3)
order by 3 limit 3)
union all
(select * from functional.alltypestiny where year=2009 and month=3
union all
select * from functional.alltypestiny where year=2009 and month=4)
union all
(select * from functional.alltypestiny where year=2009 and month=4
union all
(select * from functional.alltypestiny where year=2009 and month=5)
order by 3 limit 3)
---- PLAN
09:UNION
|
|--08:AGGREGATE [FINALIZE]
| | group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
| |
| 00:UNION
| |
| |--07:TOP-N [LIMIT=3]
| | | order by: tinyint_col ASC
| | |
| | 04:UNION
| | |
| | |--06:SCAN HDFS [functional.alltypestiny]
| | | partitions=1/4 size=115B
| | |
| | 05:SCAN HDFS [functional.alltypestiny]
| | partitions=1/4 size=115B
| |
| |--03:SCAN HDFS [functional.alltypestiny]
| | partitions=1/4 size=115B
| |
| |--02:SCAN HDFS [functional.alltypestiny]
| | partitions=1/4 size=115B
| |
| 01:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--15:TOP-N [LIMIT=3]
| | order by: tinyint_col ASC
| |
| 12:UNION
| |
| |--14:SCAN HDFS [functional.alltypestiny]
| | partitions=0/4 size=0B
| |
| 13:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--11:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
10:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 5:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
NODE 6:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=3/090301.txt 0:115
NODE 10:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=3/090301.txt 0:115
NODE 11:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=4/090401.txt 0:115
NODE 13:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=4/090401.txt 0:115
NODE 14:
---- DISTRIBUTEDPLAN
22:EXCHANGE [UNPARTITIONED]
|
09:UNION
|
|--11:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--10:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--21:EXCHANGE [RANDOM]
| |
| 16:MERGING-EXCHANGE [UNPARTITIONED]
| | order by: tinyint_col ASC
| | limit: 3
| |
| 15:TOP-N [LIMIT=3]
| | order by: tinyint_col ASC
| |
| 12:UNION
| |
| |--14:SCAN HDFS [functional.alltypestiny]
| | partitions=0/4 size=0B
| |
| 13:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
20:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
19:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
08:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
|--01:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
18:EXCHANGE [RANDOM]
|
17:MERGING-EXCHANGE [UNPARTITIONED]
| order by: tinyint_col ASC
| limit: 3
|
07:TOP-N [LIMIT=3]
| order by: tinyint_col ASC
|
04:UNION
|
|--06:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
05:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// UNION ALL in subquery
select x.* from
(select * from functional.alltypestiny where year=2009 and month=1
union all
select * from functional.alltypestiny where year=2009 and month=1) x
union all
(select * from functional.alltypestiny where year=2009 and month=2)
order by 3 limit 5
---- PLAN
05:TOP-N [LIMIT=5]
| order by: tinyint_col ASC
|
00:UNION
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
02:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 4:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
06:MERGING-EXCHANGE [UNPARTITIONED]
| order by: tinyint_col ASC
| limit: 5
|
05:TOP-N [LIMIT=5]
| order by: tinyint_col ASC
|
00:UNION
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
01:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
02:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// UNION DISTINCT in subquery
select x.* from
(select * from functional.alltypestiny where year=2009 and month=1
union distinct
select * from functional.alltypestiny where year=2009 and month=1) x
union distinct
(select * from functional.alltypestiny where year=2009 and month=2)
order by 3 limit 3
---- PLAN
07:TOP-N [LIMIT=3]
| order by: tinyint_col ASC
|
06:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--05:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
04:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
01:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
02:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- SCANRANGELOCATIONS
NODE 2:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 5:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=2/090201.txt 0:115
---- DISTRIBUTEDPLAN
12:MERGING-EXCHANGE [UNPARTITIONED]
| order by: tinyint_col ASC
| limit: 3
|
07:TOP-N [LIMIT=3]
| order by: tinyint_col ASC
|
11:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
10:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
06:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
|
|--05:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
09:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
08:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
04:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
01:UNION
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
02:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// UNION ALL in subquery with a WHERE condition in the outer select.
select x.* from
(select int_col, bool_col, count(*) as count_col
from functional.alltypestiny where year=2009 and month=1 group by 1, 2
union all
select int_col, bool_col, count(*) as count_col
from functional.alltypestiny where year=2009 and month=1 group by 1, 2) x
where x.int_col < 5 and x.bool_col = false
---- PLAN
00:UNION
|
|--04:AGGREGATE [FINALIZE]
| | output: count(*)
| | group by: int_col, bool_col
| |
| 03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
| predicates: functional.alltypestiny.int_col < 5, functional.alltypestiny.bool_col = FALSE
|
02:AGGREGATE [FINALIZE]
| output: count(*)
| group by: int_col, bool_col
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
predicates: functional.alltypestiny.int_col < 5, functional.alltypestiny.bool_col = FALSE
---- SCANRANGELOCATIONS
NODE 1:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
NODE 3:
HDFS SPLIT hdfs://localhost:20500/test-warehouse/alltypestiny/year=2009/month=1/090101.txt 0:115
---- DISTRIBUTEDPLAN
09:EXCHANGE [UNPARTITIONED]
|
00:UNION
|
|--08:AGGREGATE [MERGE FINALIZE]
| | output: sum(count(*))
| | group by: int_col, bool_col
| |
| 07:EXCHANGE [HASH(int_col,bool_col)]
| |
| 04:AGGREGATE
| | output: count(*)
| | group by: int_col, bool_col
| |
| 03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
| predicates: functional.alltypestiny.int_col < 5, functional.alltypestiny.bool_col = FALSE
|
06:AGGREGATE [MERGE FINALIZE]
| output: sum(count(*))
| group by: int_col, bool_col
|
05:EXCHANGE [HASH(int_col,bool_col)]
|
02:AGGREGATE
| output: count(*)
| group by: int_col, bool_col
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
predicates: functional.alltypestiny.int_col < 5, functional.alltypestiny.bool_col = FALSE
====
// UNION ALL with only constant selects
select 1, 'a', NULL, 10.f
union all
select 2, 'b', NULL, 20.f
union all
select 3, 'c', NULL, 30.f
---- PLAN
00:UNION
constant-operands=3
---- DISTRIBUTEDPLAN
00:UNION
constant-operands=3
====
// UNION DISTINCT with only constant selects
select 1, 'a', NULL, 10.0f
union distinct
select 2, 'b', NULL, 20.0f
union distinct
select 3, 'c', NULL, 30.0f
---- PLAN
01:AGGREGATE [FINALIZE]
| group by: 1, 'a', null, f
|
00:UNION
constant-operands=3
---- DISTRIBUTEDPLAN
01:AGGREGATE [FINALIZE]
| group by: 1, 'a', null, f
|
00:UNION
constant-operands=3
====
// UNION ALL/DISTINCT with mixed constant selects and non-constant selects
select 1, 'a', NULL, 10.f
union all
select int_col, string_col, bool_col, float_col from functional.alltypestiny
union distinct
select 3, 'c', NULL, 30.f
union all
select int_col, string_col, bool_col, float_col from functional.alltypestiny
---- PLAN
03:UNION
|
|--02:AGGREGATE [FINALIZE]
| | group by: 1, 'a', null, f
| |
| 00:UNION
| | constant-operands=2
| |
| 01:SCAN HDFS [functional.alltypestiny]
| partitions=4/4 size=460B
|
04:SCAN HDFS [functional.alltypestiny]
partitions=4/4 size=460B
---- DISTRIBUTEDPLAN
07:EXCHANGE [UNPARTITIONED]
|
03:UNION
|
|--04:SCAN HDFS [functional.alltypestiny]
| partitions=4/4 size=460B
|
06:AGGREGATE [MERGE FINALIZE]
| group by: 1, 'a', null, f
|
05:EXCHANGE [HASH(1,'a',null,f)]
|
02:AGGREGATE
| group by: 1, 'a', null, f
|
00:UNION
| constant-operands=2
|
01:SCAN HDFS [functional.alltypestiny]
partitions=4/4 size=460B
====
// UNION ALL/DISTINCT with mixed constant selects and non-constant selects and nested unions
(select 500
union all
(select int_col from functional.alltypestiny where year=2009 and month=2
order by 1 limit 3
union all
select 500)
)
union distinct
select int_col from functional.alltypestiny where year=2009 and month=1
union all
select 503
---- PLAN
05:UNION
| constant-operands=1
|
04:AGGREGATE [FINALIZE]
| group by: 500
|
00:UNION
| constant-operands=2
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
02:TOP-N [LIMIT=3]
| order by: int_col ASC
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
---- DISTRIBUTEDPLAN
10:EXCHANGE [UNPARTITIONED]
|
05:UNION
| constant-operands=1
|
09:AGGREGATE [MERGE FINALIZE]
| group by: 500
|
08:EXCHANGE [HASH(500)]
|
04:AGGREGATE
| group by: 500
|
00:UNION
| constant-operands=2
|
|--03:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
|
07:EXCHANGE [RANDOM]
|
06:MERGING-EXCHANGE [UNPARTITIONED]
| order by: int_col ASC
| limit: 3
|
02:TOP-N [LIMIT=3]
| order by: int_col ASC
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
====
// UNION ALL with only values statements
values(1, 'a', NULL, 10.f)
union all
values(2, 'b', NULL, 20.f)
union all
values(3, 'c', NULL, 30.f)
---- PLAN
00:UNION
constant-operands=3
---- DISTRIBUTEDPLAN
00:UNION
constant-operands=3
====
// UNION DISTINCT with only values statements
values(1, 'a', NULL, 10.f)
union distinct
values(2, 'b', NULL, 20.f)
union distinct
values(3, 'c', NULL, 30.f)
---- PLAN
01:AGGREGATE [FINALIZE]
| group by: 1, 'a', null, f
|
00:UNION
constant-operands=3
---- DISTRIBUTEDPLAN
01:AGGREGATE [FINALIZE]
| group by: 1, 'a', null, f
|
00:UNION
constant-operands=3
====
// UNION ALL/DISTINCT with mixed values statements and non-constant selects
values(1, 'a', NULL, 10.f)
union all
select int_col, string_col, bool_col, float_col from functional.alltypestiny
union distinct
values(3, 'c', NULL, 30.f)
union all
select int_col, string_col, bool_col, float_col from functional.alltypessmall
---- PLAN
03:UNION
|
|--02:AGGREGATE [FINALIZE]
| | group by: 1, 'a', null, f
| |
| 00:UNION
| | constant-operands=2
| |
| 01:SCAN HDFS [functional.alltypestiny]
| partitions=4/4 size=460B
|
04:SCAN HDFS [functional.alltypessmall]
partitions=4/4 size=6.32KB
---- DISTRIBUTEDPLAN
07:EXCHANGE [UNPARTITIONED]
|
03:UNION
|
|--04:SCAN HDFS [functional.alltypessmall]
| partitions=4/4 size=6.32KB
|
06:AGGREGATE [MERGE FINALIZE]
| group by: 1, 'a', null, f
|
05:EXCHANGE [HASH(1,'a',null,f)]
|
02:AGGREGATE
| group by: 1, 'a', null, f
|
00:UNION
| constant-operands=2
|
01:SCAN HDFS [functional.alltypestiny]
partitions=4/4 size=460B
====
# all union output slots are being materialized even though none is referenced by
# the enclosing scope
select count(*) from (
select id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col,
double_col, date_string_col, string_col, timestamp_col, year, month
from functional.alltypes
union distinct
select 0,true,0,0,0,0,cast(0 as float),0,'01/01/09','0',
cast('2009-01-01 00:00:00' as timestamp),2009,1
union distinct
select id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col,
double_col, date_string_col, string_col, timestamp_col, year, month
from functional.alltypes
union distinct
select 1,false,1,1,1,10,cast(1.1 as float),10.1,'01/01/09','1',
cast('2009-01-01 00:01:00' as timestamp),2009,1
union distinct
select 2,true,2,2,2,20,cast(2.2 as float),20.2,'01/01/09','2',
cast('2009-01-01 00:02:00.10' as timestamp),2009,1
) x
---- PLAN
04:AGGREGATE [FINALIZE]
| output: count(*)
|
03:AGGREGATE [FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
| constant-operands=3
|
|--02:SCAN HDFS [functional.alltypes]
| partitions=24/24 size=478.45KB
|
01:SCAN HDFS [functional.alltypes]
partitions=24/24 size=478.45KB
---- DISTRIBUTEDPLAN
08:AGGREGATE [MERGE FINALIZE]
| output: sum(count(*))
|
07:EXCHANGE [UNPARTITIONED]
|
04:AGGREGATE
| output: count(*)
|
06:AGGREGATE [MERGE FINALIZE]
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
05:EXCHANGE [HASH(id,bool_col,tinyint_col,smallint_col,int_col,bigint_col,float_col,double_col,date_string_col,string_col,timestamp_col,year,month)]
|
03:AGGREGATE
| group by: id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col, double_col, date_string_col, string_col, timestamp_col, year, month
|
00:UNION
| constant-operands=3
|
|--02:SCAN HDFS [functional.alltypes]
| partitions=24/24 size=478.45KB
|
01:SCAN HDFS [functional.alltypes]
partitions=24/24 size=478.45KB
====
# UNION ALL in subquery with a WHERE condition in the outer select;
# Where clause conjuncts are used as scan predicates and for partition pruning.
select x.int_col, x.bool_col, x.month from
(select * from functional.alltypestiny where year=2009
union all
select * from functional.alltypestiny where year=2009) x
where x.int_col < 5 and x.bool_col = false and x.month = 1
---- PLAN
00:UNION
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
| predicates: functional.alltypestiny.int_col < 5, functional.alltypestiny.bool_col = FALSE
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
predicates: functional.alltypestiny.int_col < 5, functional.alltypestiny.bool_col = FALSE
---- DISTRIBUTEDPLAN
03:EXCHANGE [UNPARTITIONED]
|
00:UNION
|
|--02:SCAN HDFS [functional.alltypestiny]
| partitions=1/4 size=115B
| predicates: functional.alltypestiny.int_col < 5, functional.alltypestiny.bool_col = FALSE
|
01:SCAN HDFS [functional.alltypestiny]
partitions=1/4 size=115B
predicates: functional.alltypestiny.int_col < 5, functional.alltypestiny.bool_col = FALSE
====
# Regression test for IMPALA-888. Tests proper materialization of slots in
# the presence of conjuncts that become constant for union operands.
select 1 from
(select 1 a, 2 b, int_col c, bigint_col d from functional.alltypestiny
union all
select tinyint_col, int_col, smallint_col, bigint_col from functional.alltypes
union all
select 1, tinyint_col, 3, bigint_col from functional.alltypessmall) t
where a + b < 100 and c > 20 and d > 50
---- PLAN
00:UNION
|
|--02:SCAN HDFS [functional.alltypes]
| partitions=24/24 size=478.45KB
| predicates: functional.alltypes.tinyint_col + functional.alltypes.int_col < 100, functional.alltypes.smallint_col > 20, functional.alltypes.bigint_col > 50
|
01:SCAN HDFS [functional.alltypestiny]
partitions=4/4 size=460B
predicates: functional.alltypestiny.int_col > 20, functional.alltypestiny.bigint_col > 50
====
# Test union where all operands are dropped because of constant conjuncts.
select * from
(select 1 a, 2 b
union all
select 3, 4
union all
select 10, 20) t
where a > b
---- PLAN
00:UNION
---- DISTRIBUTEDPLAN
00:UNION
====
# Test union merging only partitioned inputs.
select id, bigint_col from functional.alltypestiny
union all
select sum(int_col), bigint_col from functional.alltypes
where year=2009 and month=2
group by bigint_col
union all
select a.id, a.bigint_col
from functional.alltypestiny a inner join functional.alltypestiny b
on (a.id = b.id)
union all
select 1000, 2000
---- PLAN
00:UNION
| constant-operands=1
|
|--06:HASH JOIN [INNER JOIN]
| | hash predicates: a.id = b.id
| |
| |--05:SCAN HDFS [functional.alltypestiny b]
| | partitions=4/4 size=460B compact
| |
| 04:SCAN HDFS [functional.alltypestiny a]
| partitions=4/4 size=460B
|
|--03:AGGREGATE [FINALIZE]
| | output: sum(int_col)
| | group by: bigint_col
| |
| 02:SCAN HDFS [functional.alltypes]
| partitions=1/24 size=18.12KB
|
01:SCAN HDFS [functional.alltypestiny]
partitions=4/4 size=460B
---- DISTRIBUTEDPLAN
10:EXCHANGE [UNPARTITIONED]
|
00:UNION
| constant-operands=1
|
|--01:SCAN HDFS [functional.alltypestiny]
| partitions=4/4 size=460B
|
|--06:HASH JOIN [INNER JOIN, BROADCAST]
| | hash predicates: a.id = b.id
| |
| |--09:EXCHANGE [BROADCAST]
| | |
| | 05:SCAN HDFS [functional.alltypestiny b]
| | partitions=4/4 size=460B
| |
| 04:SCAN HDFS [functional.alltypestiny a]
| partitions=4/4 size=460B
|
08:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(int_col))
| group by: bigint_col
|
07:EXCHANGE [HASH(bigint_col)]
|
03:AGGREGATE
| output: sum(int_col)
| group by: bigint_col
|
02:SCAN HDFS [functional.alltypes]
partitions=1/24 size=18.12KB
====
# Test union merging only unpartitioned inputs.
select count(id), sum(bigint_col) from functional.alltypes
union all
select id, bigint_col from functional.alltypessmall order by id limit 10
union all
select 1000, 2000
---- PLAN
00:UNION
| constant-operands=1
|
|--04:TOP-N [LIMIT=10]
| | order by: id ASC
| |
| 03:SCAN HDFS [functional.alltypessmall]
| partitions=4/4 size=6.32KB
|
02:AGGREGATE [FINALIZE]
| output: count(id), sum(bigint_col)
|
01:SCAN HDFS [functional.alltypes]
partitions=24/24 size=478.45KB
---- DISTRIBUTEDPLAN
00:UNION
| constant-operands=1
|
|--07:MERGING-EXCHANGE [UNPARTITIONED]
| | order by: id ASC
| | limit: 10
| |
| 04:TOP-N [LIMIT=10]
| | order by: id ASC
| |
| 03:SCAN HDFS [functional.alltypessmall]
| partitions=4/4 size=6.32KB
|
06:AGGREGATE [MERGE FINALIZE]
| output: sum(count(id)), sum(sum(bigint_col))
|
05:EXCHANGE [UNPARTITIONED]
|
02:AGGREGATE
| output: count(id), sum(bigint_col)
|
01:SCAN HDFS [functional.alltypes]
partitions=24/24 size=478.45KB
====
# Test union merging mixed unpartitioned and partitioned inputs.
select count(id), sum(bigint_col) from functional.alltypes
union all
select id, bigint_col from functional.alltypessmall order by id limit 10
union all
select id, bigint_col from functional.alltypestiny
union all
select sum(int_col), bigint_col from functional.alltypes
where year=2009 and month=2
group by bigint_col
union all
select a.id, a.bigint_col
from functional.alltypestiny a inner join functional.alltypestiny b
on (a.id = b.id)
union all
select 1000, 2000
---- PLAN
00:UNION
| constant-operands=1
|
|--10:HASH JOIN [INNER JOIN]
| | hash predicates: a.id = b.id
| |
| |--09:SCAN HDFS [functional.alltypestiny b]
| | partitions=4/4 size=460B compact
| |
| 08:SCAN HDFS [functional.alltypestiny a]
| partitions=4/4 size=460B
|
|--07:AGGREGATE [FINALIZE]
| | output: sum(int_col)
| | group by: bigint_col
| |
| 06:SCAN HDFS [functional.alltypes]
| partitions=1/24 size=18.12KB
|
|--05:SCAN HDFS [functional.alltypestiny]
| partitions=4/4 size=460B
|
|--04:TOP-N [LIMIT=10]
| | order by: id ASC
| |
| 03:SCAN HDFS [functional.alltypessmall]
| partitions=4/4 size=6.32KB
|
02:AGGREGATE [FINALIZE]
| output: count(id), sum(bigint_col)
|
01:SCAN HDFS [functional.alltypes]
partitions=24/24 size=478.45KB
---- DISTRIBUTEDPLAN
19:EXCHANGE [UNPARTITIONED]
|
00:UNION
| constant-operands=1
|
|--05:SCAN HDFS [functional.alltypestiny]
| partitions=4/4 size=460B
|
|--18:EXCHANGE [RANDOM]
| |
| 13:MERGING-EXCHANGE [UNPARTITIONED]
| | order by: id ASC
| | limit: 10
| |
| 04:TOP-N [LIMIT=10]
| | order by: id ASC
| |
| 03:SCAN HDFS [functional.alltypessmall]
| partitions=4/4 size=6.32KB
|
|--17:EXCHANGE [RANDOM]
| |
| 12:AGGREGATE [MERGE FINALIZE]
| | output: sum(count(id)), sum(sum(bigint_col))
| |
| 11:EXCHANGE [UNPARTITIONED]
| |
| 02:AGGREGATE
| | output: count(id), sum(bigint_col)
| |
| 01:SCAN HDFS [functional.alltypes]
| partitions=24/24 size=478.45KB
|
|--10:HASH JOIN [INNER JOIN, BROADCAST]
| | hash predicates: a.id = b.id
| |
| |--16:EXCHANGE [BROADCAST]
| | |
| | 09:SCAN HDFS [functional.alltypestiny b]
| | partitions=4/4 size=460B
| |
| 08:SCAN HDFS [functional.alltypestiny a]
| partitions=4/4 size=460B
|
15:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(int_col))
| group by: bigint_col
|
14:EXCHANGE [HASH(bigint_col)]
|
07:AGGREGATE
| output: sum(int_col)
| group by: bigint_col
|
06:SCAN HDFS [functional.alltypes]
partitions=1/24 size=18.12KB
====
Reference in New Issue View Git Blame Copy Permalink