mirror of
https://github.com/apache/impala.git
synced 2026-01-10 09:00:16 -05:00
4b748ef5da
present This commit fixes the issue where predicates are not applied to the correct query tree nodes when the query contains full outer joins. To address this issue, we register information about the tuple ids that are outer joined by full outer joins and use that information to guide the assignment of predicates. Change-Id: I854c05c159d86c0aaabfc12b7dd5c5982c5ece4b Reviewed-on: http://gerrit.sjc.cloudera.com:8080/5284 Reviewed-by: Dimitris Tsirogiannis <dtsirogiannis@cloudera.com> Tested-by: jenkins
604 lines
20 KiB
Plaintext
604 lines
20 KiB
Plaintext
# correct placement of predicates with left outer joins; t2 and t3 are nullable
|
|
select *
|
|
from functional.testtbl t1 left outer join functional.testtbl t2 on (
|
|
t1.id - 1 = t2.id + 1
|
|
# lhs predicate is join predicate
|
|
and t1.zip = 94611
|
|
# rhs predicate applied by rhs scan
|
|
and t2.zip = 94104)
|
|
left outer join functional.testtbl t3 on (
|
|
t1.id = t3.id
|
|
# predicate on t2 is join predicate, not applied by t2 scan
|
|
and t2.id = 15
|
|
# predicate on t1 and t2 is join predicate
|
|
and t1.id - t2.id = 0
|
|
# rhs predicate applied by rhs scan
|
|
and t3.zip = 94720
|
|
)
|
|
where
|
|
# t1 predicate in where clause is applied by scans (t1; also propagated to t3)
|
|
t1.id > 0
|
|
# t2 predicate in where clause is applied by join and scan
|
|
and t2.id is null
|
|
# t3 predicate in where clause is applied by join and scan
|
|
and t3.id is not null
|
|
# join predicate between t1 and t2 applied after t2 join
|
|
and t1.zip + t2.zip = 10
|
|
# join predicate between t1, t2 and t3 applied after last join
|
|
and t1.zip + t2.zip + t3.zip= 20
|
|
---- PLAN
|
|
04:HASH JOIN [LEFT OUTER JOIN]
|
|
| hash predicates: t1.id = t3.id
|
|
| other join predicates: t2.id = 15, t1.id - t2.id = 0
|
|
| other predicates: t3.id IS NOT NULL, t1.zip + t2.zip + t3.zip = 20
|
|
|
|
|
|--02:SCAN HDFS [functional.testtbl t3]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: t3.zip = 94720, t3.id > 0, t3.id IS NOT NULL
|
|
|
|
|
03:HASH JOIN [LEFT OUTER JOIN]
|
|
| hash predicates: t1.id - 1 = t2.id + 1
|
|
| other join predicates: t1.zip = 94611
|
|
| other predicates: t2.id IS NULL, t1.zip + t2.zip = 10
|
|
|
|
|
|--01:SCAN HDFS [functional.testtbl t2]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: t2.zip = 94104
|
|
|
|
|
00:SCAN HDFS [functional.testtbl t1]
|
|
partitions=1/1 files=0 size=0B
|
|
predicates: t1.id > 0
|
|
---- DISTRIBUTEDPLAN
|
|
07:EXCHANGE [UNPARTITIONED]
|
|
|
|
|
04:HASH JOIN [LEFT OUTER JOIN, BROADCAST]
|
|
| hash predicates: t1.id = t3.id
|
|
| other join predicates: t2.id = 15, t1.id - t2.id = 0
|
|
| other predicates: t3.id IS NOT NULL, t1.zip + t2.zip + t3.zip = 20
|
|
|
|
|
|--06:EXCHANGE [BROADCAST]
|
|
| |
|
|
| 02:SCAN HDFS [functional.testtbl t3]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: t3.zip = 94720, t3.id > 0, t3.id IS NOT NULL
|
|
|
|
|
03:HASH JOIN [LEFT OUTER JOIN, BROADCAST]
|
|
| hash predicates: t1.id - 1 = t2.id + 1
|
|
| other join predicates: t1.zip = 94611
|
|
| other predicates: t2.id IS NULL, t1.zip + t2.zip = 10
|
|
|
|
|
|--05:EXCHANGE [BROADCAST]
|
|
| |
|
|
| 01:SCAN HDFS [functional.testtbl t2]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: t2.zip = 94104
|
|
|
|
|
00:SCAN HDFS [functional.testtbl t1]
|
|
partitions=1/1 files=0 size=0B
|
|
predicates: t1.id > 0
|
|
====
|
|
# the same thing with subqueries; should produce the same result
|
|
select *
|
|
from (select * from functional.testtbl a1) t1
|
|
left outer join (select * from functional.testtbl a2) t2 on (
|
|
t1.id - 1 = t2.id + 1 and t1.zip = 94611 and t2.zip = 94104)
|
|
left outer join (select * from functional.testtbl a3) t3 on (
|
|
t1.id = t3.id and t2.id = 15 and t1.id - t2.id = 0 and t3.zip = 94720)
|
|
where t1.id > 0 and t2.id is null and t3.id is not null
|
|
and t1.zip + t2.zip = 10 and t1.zip + t2.zip + t3.zip= 20
|
|
---- PLAN
|
|
04:HASH JOIN [LEFT OUTER JOIN]
|
|
| hash predicates: a1.id = a3.id
|
|
| other join predicates: a2.id = 15, a1.id - a2.id = 0
|
|
| other predicates: a3.id IS NOT NULL, a1.zip + a2.zip + a3.zip = 20
|
|
|
|
|
|--02:SCAN HDFS [functional.testtbl a3]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: a3.id > 0, a3.id IS NOT NULL, a3.zip = 94720
|
|
|
|
|
03:HASH JOIN [LEFT OUTER JOIN]
|
|
| hash predicates: a1.id - 1 = a2.id + 1
|
|
| other join predicates: a1.zip = 94611
|
|
| other predicates: a2.id IS NULL, a1.zip + a2.zip = 10
|
|
|
|
|
|--01:SCAN HDFS [functional.testtbl a2]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: a2.zip = 94104
|
|
|
|
|
00:SCAN HDFS [functional.testtbl a1]
|
|
partitions=1/1 files=0 size=0B
|
|
predicates: a1.id > 0
|
|
---- DISTRIBUTEDPLAN
|
|
07:EXCHANGE [UNPARTITIONED]
|
|
|
|
|
04:HASH JOIN [LEFT OUTER JOIN, BROADCAST]
|
|
| hash predicates: a1.id = a3.id
|
|
| other join predicates: a2.id = 15, a1.id - a2.id = 0
|
|
| other predicates: a3.id IS NOT NULL, a1.zip + a2.zip + a3.zip = 20
|
|
|
|
|
|--06:EXCHANGE [BROADCAST]
|
|
| |
|
|
| 02:SCAN HDFS [functional.testtbl a3]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: a3.id > 0, a3.id IS NOT NULL, a3.zip = 94720
|
|
|
|
|
03:HASH JOIN [LEFT OUTER JOIN, BROADCAST]
|
|
| hash predicates: a1.id - 1 = a2.id + 1
|
|
| other join predicates: a1.zip = 94611
|
|
| other predicates: a2.id IS NULL, a1.zip + a2.zip = 10
|
|
|
|
|
|--05:EXCHANGE [BROADCAST]
|
|
| |
|
|
| 01:SCAN HDFS [functional.testtbl a2]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: a2.zip = 94104
|
|
|
|
|
00:SCAN HDFS [functional.testtbl a1]
|
|
partitions=1/1 files=0 size=0B
|
|
predicates: a1.id > 0
|
|
====
|
|
# correct propagation of scan predicates in OJ On clauses:
|
|
# id = 17 must not be applied by the t1 and t3 scans
|
|
select *
|
|
from functional.testtbl t1
|
|
left outer join functional.testtbl t2 on (
|
|
t1.id = t2.id and t1.id = 17)
|
|
join functional.testtbl t3 on (t1.id = t3.id)
|
|
---- PLAN
|
|
04:HASH JOIN [INNER JOIN]
|
|
| hash predicates: t1.id = t3.id
|
|
|
|
|
|--02:SCAN HDFS [functional.testtbl t3]
|
|
| partitions=1/1 files=0 size=0B
|
|
|
|
|
03:HASH JOIN [LEFT OUTER JOIN]
|
|
| hash predicates: t1.id = t2.id
|
|
| other join predicates: t1.id = 17
|
|
|
|
|
|--01:SCAN HDFS [functional.testtbl t2]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: t2.id = 17
|
|
|
|
|
00:SCAN HDFS [functional.testtbl t1]
|
|
partitions=1/1 files=0 size=0B
|
|
====
|
|
# correct placement of predicates with right outer joins; t1 and t2 are nullable
|
|
select *
|
|
from functional.testtbl t1 right outer join functional.testtbl t2 on (
|
|
t1.id - 1 = t2.id + 1
|
|
# lhs predicate is scan predicate
|
|
and t1.zip = 94611
|
|
# rhs predicate is join predicate
|
|
and t2.zip = 94104)
|
|
right outer join functional.testtbl t3 on (
|
|
t1.id = t3.id
|
|
# predicate on t2 is scan predicate
|
|
and t2.id = 15
|
|
# predicate on t1 and t2 could be applied by previous join
|
|
# but the planner doesn't recognize that case and makes it a join predicate
|
|
and t1.id - t2.id = 0
|
|
# rhs predicate is join predicate
|
|
and t3.zip = 94720
|
|
)
|
|
where
|
|
# t1 predicate in where clause is applied by scan and last join
|
|
t1.id > 0
|
|
# t2 predicate in where clause is applied by last join
|
|
and t2.id is null
|
|
# t3 predicate in where clause is applied by scan
|
|
and t3.id is not null
|
|
# join predicate between t1 and t2 applied after last join
|
|
and t1.zip + t2.zip = 10
|
|
# join predicate between t1, t2 and t3 applied after last join
|
|
and t1.zip + t2.zip + t3.zip= 20
|
|
---- PLAN
|
|
04:HASH JOIN [RIGHT OUTER JOIN]
|
|
| hash predicates: t1.id = t3.id
|
|
| other join predicates: t1.id - t2.id = 0, t3.zip = 94720
|
|
| other predicates: t1.id > 0, t2.id IS NULL, t1.zip + t2.zip = 10, t1.zip + t2.zip + t3.zip = 20
|
|
|
|
|
|--02:SCAN HDFS [functional.testtbl t3]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: t3.id IS NOT NULL
|
|
|
|
|
03:HASH JOIN [RIGHT OUTER JOIN]
|
|
| hash predicates: t1.id - 1 = t2.id + 1
|
|
| other join predicates: t1.zip = 94611, t2.zip = 94104
|
|
|
|
|
|--01:SCAN HDFS [functional.testtbl t2]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: t2.id = 15
|
|
|
|
|
00:SCAN HDFS [functional.testtbl t1]
|
|
partitions=1/1 files=0 size=0B
|
|
predicates: t1.id > 0, t1.id IS NOT NULL
|
|
---- DISTRIBUTEDPLAN
|
|
09:EXCHANGE [UNPARTITIONED]
|
|
|
|
|
04:HASH JOIN [RIGHT OUTER JOIN, PARTITIONED]
|
|
| hash predicates: t1.id = t3.id
|
|
| other join predicates: t1.id - t2.id = 0, t3.zip = 94720
|
|
| other predicates: t1.id > 0, t2.id IS NULL, t1.zip + t2.zip = 10, t1.zip + t2.zip + t3.zip = 20
|
|
|
|
|
|--08:EXCHANGE [HASH(t3.id)]
|
|
| |
|
|
| 02:SCAN HDFS [functional.testtbl t3]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: t3.id IS NOT NULL
|
|
|
|
|
07:EXCHANGE [HASH(t1.id)]
|
|
|
|
|
03:HASH JOIN [RIGHT OUTER JOIN, PARTITIONED]
|
|
| hash predicates: t1.id - 1 = t2.id + 1
|
|
| other join predicates: t1.zip = 94611, t2.zip = 94104
|
|
|
|
|
|--06:EXCHANGE [HASH(t2.id + 1)]
|
|
| |
|
|
| 01:SCAN HDFS [functional.testtbl t2]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: t2.id = 15
|
|
|
|
|
05:EXCHANGE [HASH(t1.id - 1)]
|
|
|
|
|
00:SCAN HDFS [functional.testtbl t1]
|
|
partitions=1/1 files=0 size=0B
|
|
predicates: t1.id > 0, t1.id IS NOT NULL
|
|
====
|
|
# the same thing with subqueries; should produce the same result
|
|
select *
|
|
from (select * from functional.testtbl a1) t1 right outer join (select * from functional.testtbl a2) t2 on (
|
|
t1.id - 1 = t2.id + 1 and t1.zip = 94611 and t2.zip = 94104)
|
|
right outer join (select * from functional.testtbl a3) t3 on (
|
|
t1.id = t3.id and t2.id = 15 and t1.id - t2.id = 0 and t3.zip = 94720 )
|
|
where t1.id > 0 and t2.id is null and t3.id is not null
|
|
and t1.zip + t2.zip = 10 and t1.zip + t2.zip + t3.zip= 20
|
|
---- PLAN
|
|
04:HASH JOIN [RIGHT OUTER JOIN]
|
|
| hash predicates: a1.id = a3.id
|
|
| other join predicates: a1.id - a2.id = 0, a3.zip = 94720
|
|
| other predicates: a2.id IS NULL, a1.id > 0, a1.zip + a2.zip = 10, a1.zip + a2.zip + a3.zip = 20
|
|
|
|
|
|--02:SCAN HDFS [functional.testtbl a3]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: a3.id IS NOT NULL
|
|
|
|
|
03:HASH JOIN [RIGHT OUTER JOIN]
|
|
| hash predicates: a1.id - 1 = a2.id + 1
|
|
| other join predicates: a1.zip = 94611, a2.zip = 94104
|
|
|
|
|
|--01:SCAN HDFS [functional.testtbl a2]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: a2.id = 15
|
|
|
|
|
00:SCAN HDFS [functional.testtbl a1]
|
|
partitions=1/1 files=0 size=0B
|
|
predicates: a1.id > 0, a1.id IS NOT NULL
|
|
---- DISTRIBUTEDPLAN
|
|
09:EXCHANGE [UNPARTITIONED]
|
|
|
|
|
04:HASH JOIN [RIGHT OUTER JOIN, PARTITIONED]
|
|
| hash predicates: a1.id = a3.id
|
|
| other join predicates: a1.id - a2.id = 0, a3.zip = 94720
|
|
| other predicates: a2.id IS NULL, a1.id > 0, a1.zip + a2.zip = 10, a1.zip + a2.zip + a3.zip = 20
|
|
|
|
|
|--08:EXCHANGE [HASH(a3.id)]
|
|
| |
|
|
| 02:SCAN HDFS [functional.testtbl a3]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: a3.id IS NOT NULL
|
|
|
|
|
07:EXCHANGE [HASH(a1.id)]
|
|
|
|
|
03:HASH JOIN [RIGHT OUTER JOIN, PARTITIONED]
|
|
| hash predicates: a1.id - 1 = a2.id + 1
|
|
| other join predicates: a1.zip = 94611, a2.zip = 94104
|
|
|
|
|
|--06:EXCHANGE [HASH(a2.id + 1)]
|
|
| |
|
|
| 01:SCAN HDFS [functional.testtbl a2]
|
|
| partitions=1/1 files=0 size=0B
|
|
| predicates: a2.id = 15
|
|
|
|
|
05:EXCHANGE [HASH(a1.id - 1)]
|
|
|
|
|
00:SCAN HDFS [functional.testtbl a1]
|
|
partitions=1/1 files=0 size=0B
|
|
predicates: a1.id > 0, a1.id IS NOT NULL
|
|
====
|
|
# right outer join requires the join op to be partitioned, otherwise non-matches cause
|
|
# duplicates
|
|
select a.tinyint_col, b.id
|
|
from functional.alltypesagg a
|
|
right outer join functional.alltypestiny b on (a.tinyint_col = b.id)
|
|
where a.tinyint_col is null
|
|
---- PLAN
|
|
02:HASH JOIN [RIGHT OUTER JOIN]
|
|
| hash predicates: a.tinyint_col = b.id
|
|
| other predicates: a.tinyint_col IS NULL
|
|
|
|
|
|--01:SCAN HDFS [functional.alltypestiny b]
|
|
| partitions=4/4 files=4 size=460B
|
|
|
|
|
00:SCAN HDFS [functional.alltypesagg a]
|
|
partitions=11/11 files=11 size=814.73KB
|
|
---- DISTRIBUTEDPLAN
|
|
05:EXCHANGE [UNPARTITIONED]
|
|
|
|
|
02:HASH JOIN [RIGHT OUTER JOIN, PARTITIONED]
|
|
| hash predicates: a.tinyint_col = b.id
|
|
| other predicates: a.tinyint_col IS NULL
|
|
|
|
|
|--04:EXCHANGE [HASH(b.id)]
|
|
| |
|
|
| 01:SCAN HDFS [functional.alltypestiny b]
|
|
| partitions=4/4 files=4 size=460B
|
|
|
|
|
03:EXCHANGE [HASH(a.tinyint_col)]
|
|
|
|
|
00:SCAN HDFS [functional.alltypesagg a]
|
|
partitions=11/11 files=11 size=814.73KB
|
|
====
|
|
# same for full outer joins
|
|
select a.tinyint_col, b.id
|
|
from functional.alltypesagg a
|
|
full outer join functional.alltypestiny b on (a.tinyint_col = b.id)
|
|
where a.tinyint_col is null
|
|
---- PLAN
|
|
02:HASH JOIN [FULL OUTER JOIN]
|
|
| hash predicates: a.tinyint_col = b.id
|
|
| other predicates: a.tinyint_col IS NULL
|
|
|
|
|
|--01:SCAN HDFS [functional.alltypestiny b]
|
|
| partitions=4/4 files=4 size=460B
|
|
|
|
|
00:SCAN HDFS [functional.alltypesagg a]
|
|
partitions=11/11 files=11 size=814.73KB
|
|
---- DISTRIBUTEDPLAN
|
|
05:EXCHANGE [UNPARTITIONED]
|
|
|
|
|
02:HASH JOIN [FULL OUTER JOIN, PARTITIONED]
|
|
| hash predicates: a.tinyint_col = b.id
|
|
| other predicates: a.tinyint_col IS NULL
|
|
|
|
|
|--04:EXCHANGE [HASH(b.id)]
|
|
| |
|
|
| 01:SCAN HDFS [functional.alltypestiny b]
|
|
| partitions=4/4 files=4 size=460B
|
|
|
|
|
03:EXCHANGE [HASH(a.tinyint_col)]
|
|
|
|
|
00:SCAN HDFS [functional.alltypesagg a]
|
|
partitions=11/11 files=11 size=814.73KB
|
|
====
|
|
# Predicate assignment when query has a full outer join (IMPALA-1371)
|
|
select *
|
|
from functional.alltypes a full outer join functional.alltypes b
|
|
# predicate on a from the on clause is not assigned to the scan node
|
|
on (a.id = b.id and a.int_col < 10 and b.tinyint_col != 5)
|
|
# predicate on b from the where clause is assigned to the scan and
|
|
# also to the full outer join
|
|
where b.bigint_col > 10
|
|
---- PLAN
|
|
02:HASH JOIN [FULL OUTER JOIN]
|
|
| hash predicates: a.id = b.id
|
|
| other join predicates: a.int_col < 10, b.tinyint_col != 5
|
|
| other predicates: b.bigint_col > 10
|
|
|
|
|
|--01:SCAN HDFS [functional.alltypes b]
|
|
| partitions=24/24 files=24 size=478.45KB
|
|
| predicates: b.bigint_col > 10
|
|
|
|
|
00:SCAN HDFS [functional.alltypes a]
|
|
partitions=24/24 files=24 size=478.45KB
|
|
====
|
|
# Predicate assignment when query has a full outer followed by an inner join
|
|
select *
|
|
from functional.alltypes a full outer join functional.alltypes b
|
|
on (a.id = b.id)
|
|
inner join functional.alltypes c
|
|
on (a.int_col = c.int_col and a.tinyint_col < 10)
|
|
# predicate on b from the where clause is assigned to the
|
|
# first full outer join
|
|
where b.tinyint_col > 20
|
|
---- PLAN
|
|
05:HASH JOIN [INNER JOIN]
|
|
| hash predicates: a.int_col = c.int_col
|
|
|
|
|
|--02:SCAN HDFS [functional.alltypes c]
|
|
| partitions=24/24 files=24 size=478.45KB
|
|
|
|
|
04:HASH JOIN [FULL OUTER JOIN]
|
|
| hash predicates: a.id = b.id
|
|
| other predicates: a.tinyint_col < 10, b.tinyint_col > 20
|
|
|
|
|
|--01:SCAN HDFS [functional.alltypes b]
|
|
| partitions=24/24 files=24 size=478.45KB
|
|
| predicates: b.tinyint_col > 20
|
|
|
|
|
00:SCAN HDFS [functional.alltypes a]
|
|
partitions=24/24 files=24 size=478.45KB
|
|
predicates: a.tinyint_col < 10
|
|
====
|
|
# Predicate assignment when query has an inner join followed by a full outer join
|
|
select *
|
|
from functional.alltypes a inner join functional.alltypes b
|
|
on (a.id = b.id)
|
|
full outer join functional.alltypes c
|
|
on (a.int_col = c.int_col and a.tinyint_col < b.tinyint_col and a.bigint_col < 10)
|
|
# predicate on a is assigned to the scan and the full outer join, predicate on a and b
|
|
# is assigned to the full outer join
|
|
# TODO Ideally, the predicate on a and b should be assigned to the inner join and then
|
|
# re-assigned to the full outer join.
|
|
where a.smallint_col = 100 and a.float_col > b.float_col
|
|
---- PLAN
|
|
04:HASH JOIN [FULL OUTER JOIN]
|
|
| hash predicates: a.int_col = c.int_col
|
|
| other join predicates: a.tinyint_col < b.tinyint_col, a.bigint_col < 10
|
|
| other predicates: a.smallint_col = 100, a.float_col > b.float_col
|
|
|
|
|
|--02:SCAN HDFS [functional.alltypes c]
|
|
| partitions=24/24 files=24 size=478.45KB
|
|
|
|
|
03:HASH JOIN [INNER JOIN]
|
|
| hash predicates: b.id = a.id
|
|
|
|
|
|--00:SCAN HDFS [functional.alltypes a]
|
|
| partitions=24/24 files=24 size=478.45KB
|
|
| predicates: a.smallint_col = 100
|
|
|
|
|
01:SCAN HDFS [functional.alltypes b]
|
|
partitions=24/24 files=24 size=478.45KB
|
|
====
|
|
# Predicate assignment when query has a sequence of full outer joins interleaved with
|
|
# an inner join
|
|
select *
|
|
from functional.alltypes a full outer join functional.alltypes b
|
|
on (a.id = b.id and a.int_col < 10)
|
|
inner join functional.alltypes c
|
|
on (a.tinyint_col = c.tinyint_col and b.int_col > 10)
|
|
full outer join functional.alltypes d
|
|
# predicate on b from the on clause is assigned to the scan and is also re-applied
|
|
# in the full outer join
|
|
on (a.tinyint_col = d.tinyint_col and b.int_col < 20)
|
|
# predicate on b from the where clause is assigned to the first full outer join
|
|
where a.bool_col = false and a.float_col < b.float_col
|
|
---- PLAN
|
|
07:HASH JOIN [FULL OUTER JOIN]
|
|
| hash predicates: a.tinyint_col = d.tinyint_col
|
|
| other join predicates: b.int_col < 20
|
|
| other predicates: a.bool_col = FALSE, a.float_col < b.float_col
|
|
|
|
|
|--03:SCAN HDFS [functional.alltypes d]
|
|
| partitions=24/24 files=24 size=478.45KB
|
|
|
|
|
06:HASH JOIN [INNER JOIN]
|
|
| hash predicates: a.tinyint_col = c.tinyint_col
|
|
|
|
|
|--02:SCAN HDFS [functional.alltypes c]
|
|
| partitions=24/24 files=24 size=478.45KB
|
|
|
|
|
05:HASH JOIN [FULL OUTER JOIN]
|
|
| hash predicates: a.id = b.id
|
|
| other join predicates: a.int_col < 10
|
|
| other predicates: b.int_col > 10
|
|
|
|
|
|--01:SCAN HDFS [functional.alltypes b]
|
|
| partitions=24/24 files=24 size=478.45KB
|
|
| predicates: b.int_col > 10
|
|
|
|
|
00:SCAN HDFS [functional.alltypes a]
|
|
partitions=24/24 files=24 size=478.45KB
|
|
predicates: a.bool_col = FALSE
|
|
====
|
|
# Predicate assignment through an inline-view that has a full outer join
|
|
select 1
|
|
from (
|
|
select a.tinyint_col x, b.tinyint_col y
|
|
from functional.alltypes a
|
|
full outer join functional.alltypes b
|
|
on (a.id = b.id)
|
|
inner join functional.alltypes c
|
|
# predicate on a and b from the on clause of the inner join is assigned to
|
|
# the first full outer join
|
|
on (b.smallint_col = c.smallint_col and a.int_col < b.int_col and b.id < 10)
|
|
full outer join functional.alltypes d
|
|
# predicate on a and b from the on clause of the full outer join is assigned
|
|
# to the full outer join
|
|
on (c.id = d.id and a.bigint_col > b.bigint_col)) v
|
|
# predicate not specified in the scope of the inline view is assigned to the last
|
|
# full outer join of the inline view that materializes the corresponding
|
|
# tuple ids
|
|
where x != y
|
|
---- PLAN
|
|
06:HASH JOIN [FULL OUTER JOIN]
|
|
| hash predicates: c.id = d.id
|
|
| other join predicates: a.bigint_col > b.bigint_col
|
|
| other predicates: a.tinyint_col != b.tinyint_col
|
|
|
|
|
|--03:SCAN HDFS [functional.alltypes d]
|
|
| partitions=24/24 files=24 size=478.45KB
|
|
|
|
|
05:HASH JOIN [INNER JOIN]
|
|
| hash predicates: b.smallint_col = c.smallint_col
|
|
|
|
|
|--02:SCAN HDFS [functional.alltypes c]
|
|
| partitions=24/24 files=24 size=478.45KB
|
|
|
|
|
04:HASH JOIN [FULL OUTER JOIN]
|
|
| hash predicates: a.id = b.id
|
|
| other predicates: a.int_col < b.int_col, b.id < 10
|
|
|
|
|
|--01:SCAN HDFS [functional.alltypes b]
|
|
| partitions=24/24 files=24 size=478.45KB
|
|
| predicates: b.id < 10
|
|
|
|
|
00:SCAN HDFS [functional.alltypes a]
|
|
partitions=24/24 files=24 size=478.45KB
|
|
====
|
|
# Predicate assignment when query has an inner join followed by full outer join
|
|
select 1
|
|
from functional.alltypes a
|
|
inner join functional.alltypes b
|
|
on (a.id = b.id)
|
|
full outer join functional.alltypes c
|
|
on (b.int_col = c.int_col and c.int_col < 10)
|
|
# predicates on a and b from the where clause are assigned to the last full outer join
|
|
# that materializes the corresponding tuple ids
|
|
where a.bigint_col = b.bigint_col and a.tinyint_col < b.tinyint_col
|
|
---- PLAN
|
|
04:HASH JOIN [FULL OUTER JOIN]
|
|
| hash predicates: b.int_col = c.int_col
|
|
| other join predicates: c.int_col < 10
|
|
| other predicates: a.bigint_col = b.bigint_col, a.tinyint_col < b.tinyint_col
|
|
|
|
|
|--02:SCAN HDFS [functional.alltypes c]
|
|
| partitions=24/24 files=24 size=478.45KB
|
|
|
|
|
03:HASH JOIN [INNER JOIN]
|
|
| hash predicates: b.id = a.id
|
|
|
|
|
|--00:SCAN HDFS [functional.alltypes a]
|
|
| partitions=24/24 files=24 size=478.45KB
|
|
|
|
|
01:SCAN HDFS [functional.alltypes b]
|
|
partitions=24/24 files=24 size=478.45KB
|
|
====
|
|
# Predicate assignment when query has a sequence of inner, outer and full
|
|
# outer joins
|
|
select 1
|
|
from functional.alltypes a
|
|
inner join functional.alltypes b
|
|
on (a.id = b.id)
|
|
left outer join functional.alltypes c
|
|
on (a.int_col = c.int_col and c.tinyint_col = 10)
|
|
full outer join functional.alltypes d
|
|
on (b.string_col = d.string_col and a.tinyint_col < b.tinyint_col)
|
|
where a.float_col = b.float_col and b.smallint_col = 1 and d.tinyint_col < 10
|
|
---- PLAN
|
|
06:HASH JOIN [FULL OUTER JOIN]
|
|
| hash predicates: b.string_col = d.string_col
|
|
| other join predicates: a.tinyint_col < b.tinyint_col
|
|
| other predicates: a.float_col = b.float_col, b.smallint_col = 1, d.tinyint_col < 10
|
|
|
|
|
|--03:SCAN HDFS [functional.alltypes d]
|
|
| partitions=24/24 files=24 size=478.45KB
|
|
| predicates: d.tinyint_col < 10
|
|
|
|
|
05:HASH JOIN [LEFT OUTER JOIN]
|
|
| hash predicates: a.int_col = c.int_col
|
|
| other join predicates: c.tinyint_col = 10
|
|
|
|
|
|--02:SCAN HDFS [functional.alltypes c]
|
|
| partitions=24/24 files=24 size=478.45KB
|
|
|
|
|
04:HASH JOIN [INNER JOIN]
|
|
| hash predicates: a.id = b.id
|
|
|
|
|
|--01:SCAN HDFS [functional.alltypes b]
|
|
| partitions=24/24 files=24 size=478.45KB
|
|
| predicates: b.smallint_col = 1
|
|
|
|
|
00:SCAN HDFS [functional.alltypes a]
|
|
partitions=24/24 files=24 size=478.45KB
|
|
====
|