Hao Zhu created DRILL-1325:
------------------------------
Summary: SQL with non-equal join failed.
Key: DRILL-1325
URL: https://issues.apache.org/jira/browse/DRILL-1325
Project: Apache Drill
Issue Type: Bug
Components: Query Planning & Optimization
Affects Versions: 0.4.0
Environment: MapR 3.1.1
Drill 0.4.0
CentOS 6.4
Reporter: Hao Zhu
Priority: Critical
Equal hash join works fine, for example:
{code}
> explain plan for select a.*,b.user_port
. . . . . . . . . . . . . . . . . . . . . . .> from cp.`employee.json` a,
sys.drillbits b
. . . . . . . . . . . . . . . . . . . . . . .> where a.position_id=b.user_port
;
+------------+------------+
| text | json |
+------------+------------+
| 00-00 Screen
00-01 Project(*=[$0], user_port=[$1])
00-02 Project(T1¦¦*=[$0], user_port=[$2])
00-03 HashJoin(condition=[=($1, $2)], joinType=[inner])
00-05 |
+------------+------------+
1 row selected (0.428 seconds)
{code}
However if we change "=" to "<>" , it failed:
{code}
> explain plan for select a.*,b.user_port
. . . . . . . . . . . . . . . . . . . . . . .> from cp.`employee.json` a,
sys.drillbits b
. . . . . . . . . . . . . . . . . . . . . . .> where
a.position_id<>b.user_port ;
Query failed: Failure while parsing sql. Node
[rel#2056:Subset#4.LOGICAL.ANY([]).[]] could not be implemented; planner state:
Root: rel#2056:Subset#4.LOGICAL.ANY([]).[]
Original rel:
AbstractConverter(subset=[rel#2056:Subset#4.LOGICAL.ANY([]).[]],
convention=[LOGICAL], DrillDistributionTraitDef=[ANY([])], sort=[[]]): rowcount
= 1.7976931348623157E308, cumulative cost = {inf}, id = 2057
ProjectRel(subset=[rel#2055:Subset#4.NONE.ANY([]).[]], *=[$0],
user_port=[$3]): rowcount = 1.7976931348623157E308, cumulative cost =
{1.7976931348623157E308 rows, Infinity cpu, 0.0 io, 0.0 network}, id = 2054
FilterRel(subset=[rel#2053:Subset#3.NONE.ANY([]).[]],
condition=[<>(CAST($1):INTEGER, $3)]): rowcount = 8.988465674311579E307,
cumulative cost = {8.988465674311579E307 rows, 1.7976931348623157E308 cpu, 0.0
io, 0.0 network}, id = 2052
JoinRel(subset=[rel#2051:Subset#2.NONE.ANY([]).[]], condition=[true],
joinType=[inner]): rowcount = 10000.0, cumulative cost = {10000.0 rows, 0.0
cpu, 0.0 io, 0.0 network}, id = 2050
EnumerableTableAccessRel(subset=[rel#2048:Subset#0.ENUMERABLE.ANY([]).[]],
table=[[cp, employee.json]]): rowcount = 100.0, cumulative cost = {100.0 rows,
101.0 cpu, 0.0 io, 0.0 network}, id = 2043
EnumerableTableAccessRel(subset=[rel#2049:Subset#1.ENUMERABLE.ANY([]).[]],
table=[[sys, drillbits]]): rowcount = 100.0, cumulative cost = {100.0 rows,
101.0 cpu, 0.0 io, 0.0 network}, id = 2044
Sets:
Set#0, type: (DrillRecordRow[*, position_id])
rel#2048:Subset#0.ENUMERABLE.ANY([]).[], best=rel#2043,
importance=0.6561
rel#2043:EnumerableTableAccessRel.ENUMERABLE.ANY([]).[](table=[cp,
employee.json]), rowcount=100.0, cumulative cost={100.0 rows, 101.0 cpu, 0.0
io, 0.0 network}
rel#2093:AbstractConverter.ENUMERABLE.ANY([]).[](child=rel#2092:Subset#0.LOGICAL.ANY([]).[],convention=ENUMERABLE,DrillDistributionTraitDef=ANY([]),sort=[]),
rowcount=463.0, cumulative cost={inf}
rel#2120:AbstractConverter.ENUMERABLE.ANY([]).[](child=rel#2119:Subset#0.NONE.ANY([]).[],convention=ENUMERABLE,DrillDistributionTraitDef=ANY([]),sort=[]),
rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2092:Subset#0.LOGICAL.ANY([]).[], best=rel#2126,
importance=0.5904900000000001
rel#2094:AbstractConverter.LOGICAL.ANY([]).[](child=rel#2048:Subset#0.ENUMERABLE.ANY([]).[],convention=LOGICAL,DrillDistributionTraitDef=ANY([]),sort=[]),
rowcount=100.0, cumulative cost={inf}
rel#2121:AbstractConverter.LOGICAL.ANY([]).[](child=rel#2119:Subset#0.NONE.ANY([]).[],convention=LOGICAL,DrillDistributionTraitDef=ANY([]),sort=[]),
rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2102:AbstractConverter.LOGICAL.ANY([]).[](child=rel#2119:Subset#0.NONE.ANY([]).[],convention=LOGICAL,DrillDistributionTraitDef=ANY([]),sort=[]),
rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2118:DrillProjectRel.LOGICAL.ANY([]).[](child=rel#2092:Subset#0.LOGICAL.ANY([]).[],*=$0,position_id=$1),
rowcount=463.0, cumulative cost={926.0 rows, 934.0 cpu, 0.0 io, 0.0 network}
rel#2126:DrillScanRel.LOGICAL.ANY([]).[](table=[cp,
employee.json],groupscan=EasyGroupScan [selectionRoot=/employee.json, columns =
null]), rowcount=463.0, cumulative cost={463.0 rows, 926.0 cpu, 0.0 io, 0.0
network}
rel#2119:Subset#0.NONE.ANY([]).[], best=null, importance=0.531441
rel#2122:AbstractConverter.NONE.ANY([]).[](child=rel#2048:Subset#0.ENUMERABLE.ANY([]).[],convention=NONE,DrillDistributionTraitDef=ANY([]),sort=[]),
rowcount=100.0, cumulative cost={inf}
rel#2123:AbstractConverter.NONE.ANY([]).[](child=rel#2092:Subset#0.LOGICAL.ANY([]).[],convention=NONE,DrillDistributionTraitDef=ANY([]),sort=[]),
rowcount=463.0, cumulative cost={inf}
rel#2080:ProjectRel.NONE.ANY([]).[](child=rel#2048:Subset#0.ENUMERABLE.ANY([]).[],*=$0,position_id=$1),
rowcount=100.0, cumulative cost={inf}
Set#1, type: RecordType(VARCHAR(1) host, INTEGER user_port, INTEGER
control_port, INTEGER data_port)
rel#2049:Subset#1.ENUMERABLE.ANY([]).[], best=rel#2044,
importance=0.6561
rel#2044:EnumerableTableAccessRel.ENUMERABLE.ANY([]).[](table=[sys,
drillbits]), rowcount=100.0, cumulative cost={100.0 rows, 101.0 cpu, 0.0 io,
0.0 network}
rel#2096:AbstractConverter.ENUMERABLE.ANY([]).[](child=rel#2095:Subset#1.LOGICAL.ANY([]).[],convention=ENUMERABLE,DrillDistributionTraitDef=ANY([]),sort=[]),
rowcount=20.0, cumulative cost={inf}
rel#2095:Subset#1.LOGICAL.ANY([]).[], best=rel#2125,
importance=0.5904900000000001
rel#2097:AbstractConverter.LOGICAL.ANY([]).[](child=rel#2049:Subset#1.ENUMERABLE.ANY([]).[],convention=LOGICAL,DrillDistributionTraitDef=ANY([]),sort=[]),
rowcount=100.0, cumulative cost={inf}
rel#2125:DrillScanRel.LOGICAL.ANY([]).[](table=[sys,
drillbits],groupscan=SystemTableScan: DRILLBITS), rowcount=20.0, cumulative
cost={20.0 rows, 80.0 cpu, 0.0 io, 0.0 network}
Set#2, type: RecordType(ANY *, ANY position_id, VARCHAR(1) host, INTEGER
user_port, INTEGER control_port, INTEGER data_port)
rel#2051:Subset#2.NONE.ANY([]).[], best=null,
importance=0.7290000000000001
rel#2050:JoinRel.NONE.ANY([]).[](left=rel#2048:Subset#0.ENUMERABLE.ANY([]).[],right=rel#2049:Subset#1.ENUMERABLE.ANY([]).[],condition=true,joinType=inner),
rowcount=10000.0, cumulative cost={inf}
rel#2099:Subset#2.LOGICAL.ANY([]).[], best=null,
importance=0.7290000000000001
rel#2100:AbstractConverter.LOGICAL.ANY([]).[](child=rel#2051:Subset#2.NONE.ANY([]).[],convention=LOGICAL,DrillDistributionTraitDef=ANY([]),sort=[]),
rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2098:DrillJoinRel.LOGICAL.ANY([]).[](left=rel#2092:Subset#0.LOGICAL.ANY([]).[],right=rel#2095:Subset#1.LOGICAL.ANY([]).[],condition=true,joinType=inner),
rowcount=463.0, cumulative cost={inf}
Set#3, type: RecordType(ANY *, ANY position_id, VARCHAR(1) host, INTEGER
user_port, INTEGER control_port, INTEGER data_port)
rel#2053:Subset#3.NONE.ANY([]).[], best=null, importance=0.81
rel#2052:FilterRel.NONE.ANY([]).[](child=rel#2051:Subset#2.NONE.ANY([]).[],condition=<>(CAST($1):INTEGER,
$3)), rowcount=8.988465674311579E307, cumulative cost={inf}
rel#2079:JoinRel.NONE.ANY([]).[](left=rel#2048:Subset#0.ENUMERABLE.ANY([]).[],right=rel#2049:Subset#1.ENUMERABLE.ANY([]).[],condition=<>(CAST($1):INTEGER,
$3),joinType=inner), rowcount=5000.0, cumulative cost={inf}
rel#2068:Subset#3.LOGICAL.ANY([]).[], best=null, importance=0.81
rel#2069:AbstractConverter.LOGICAL.ANY([]).[](child=rel#2053:Subset#3.NONE.ANY([]).[],convention=LOGICAL,DrillDistributionTraitDef=ANY([]),sort=[]),
rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2109:DrillFilterRel.LOGICAL.ANY([]).[](child=rel#2099:Subset#2.LOGICAL.ANY([]).[],condition=<>(CAST($1):INTEGER,
$3)), rowcount=8.988465674311579E307, cumulative cost={inf}
Set#4, type: RecordType(ANY *, INTEGER user_port)
rel#2055:Subset#4.NONE.ANY([]).[], best=null, importance=0.9
rel#2054:ProjectRel.NONE.ANY([]).[](child=rel#2053:Subset#3.NONE.ANY([]).[],*=$0,user_port=$3),
rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2064:ProjectRel.NONE.ANY([]).[](child=rel#2063:Subset#6.NONE.ANY([]).[],*=$0,user_port=$2),
rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2056:Subset#4.LOGICAL.ANY([]).[], best=null, importance=1.0
rel#2057:AbstractConverter.LOGICAL.ANY([]).[](child=rel#2055:Subset#4.NONE.ANY([]).[],convention=LOGICAL,DrillDistributionTraitDef=ANY([]),sort=[]),
rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2067:DrillProjectRel.LOGICAL.ANY([]).[](child=rel#2065:Subset#6.LOGICAL.ANY([]).[],*=$0,user_port=$2),
rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2070:DrillProjectRel.LOGICAL.ANY([]).[](child=rel#2068:Subset#3.LOGICAL.ANY([]).[],*=$0,user_port=$3),
rowcount=1.7976931348623157E308, cumulative cost={inf}
Set#5, type: RecordType(ANY *, ANY position_id, INTEGER user_port)
rel#2061:Subset#5.NONE.ANY([]).[], best=null,
importance=0.7290000000000001
rel#2058:ProjectRel.NONE.ANY([]).[](child=rel#2051:Subset#2.NONE.ANY([]).[],*=$0,position_id=$1,user_port=$3),
rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2113:JoinRel.NONE.ANY([]).[](left=rel#2119:Subset#0.NONE.ANY([]).[],right=rel#2084:Subset#8.NONE.ANY([]).[],condition=true,joinType=inner)Error:
exception while executing query: Failure while trying to get next result
batch. (state=,code=0), rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2106:Subset#5.LOGICAL.ANY([]).[], best=null, importance=0.81
rel#2107:AbstractConverter.LOGICAL.ANY([]).[](child=rel#2061:Subset#5.NONE.ANY([]).[],convention=LOGICAL,DrillDistributionTraitDef=ANY([]),sort=[]),
rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2116:DrillJoinRel.LOGICAL.ANY([]).[](left=rel#2092:Subset#0.LOGICAL.ANY([]).[],right=rel#2103:Subset#8.LOGICAL.ANY([]).[],condition=true,joinType=inner),
rowcount=463.0, cumulative cost={inf}
rel#2124:DrillProjectRel.LOGICAL.ANY([]).[](child=rel#2099:Subset#2.LOGICAL.ANY([]).[],*=$0,position_id=$1,user_port=$3),
rowcount=1.7976931348623157E308, cumulative cost={inf}
Set#6, type: RecordType(ANY *, ANY position_id, INTEGER user_port)
rel#2063:Subset#6.NONE.ANY([]).[], best=null, importance=0.81
rel#2062:FilterRel.NONE.ANY([]).[](child=rel#2061:Subset#5.NONE.ANY([]).[],condition=<>(CAST($1):INTEGER,
$2)), rowcount=8.988465674311579E307, cumulative cost={inf}
rel#2073:ProjectRel.NONE.ANY([]).[](child=rel#2053:Subset#3.NONE.ANY([]).[],*=$0,position_id=$1,user_port=$3),
rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2085:JoinRel.NONE.ANY([]).[](left=rel#2119:Subset#0.NONE.ANY([]).[],right=rel#2084:Subset#8.NONE.ANY([]).[],condition=<>(CAST($1):INTEGER,
$2),joinType=inner), rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2065:Subset#6.LOGICAL.ANY([]).[], best=null, importance=0.9
rel#2066:AbstractConverter.LOGICAL.ANY([]).[](child=rel#2063:Subset#6.NONE.ANY([]).[],convention=LOGICAL,DrillDistributionTraitDef=ANY([]),sort=[]),
rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2105:DrillProjectRel.LOGICAL.ANY([]).[](child=rel#2068:Subset#3.LOGICAL.ANY([]).[],*=$0,position_id=$1,user_port=$3),
rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2108:DrillFilterRel.LOGICAL.ANY([]).[](child=rel#2106:Subset#5.LOGICAL.ANY([]).[],condition=<>(CAST($1):INTEGER,
$2)), rowcount=8.988465674311579E307, cumulative cost={inf}
Set#8, type: RecordType(INTEGER user_port)
rel#2084:Subset#8.NONE.ANY([]).[], best=null, importance=0.6561
rel#2081:ProjectRel.NONE.ANY([]).[](child=rel#2049:Subset#1.ENUMERABLE.ANY([]).[],user_port=$1),
rowcount=100.0, cumulative cost={inf}
rel#2103:Subset#8.LOGICAL.ANY([]).[], best=rel#2117,
importance=0.7290000000000001
rel#2104:AbstractConverter.LOGICAL.ANY([]).[](child=rel#2084:Subset#8.NONE.ANY([]).[],convention=LOGICAL,DrillDistributionTraitDef=ANY([]),sort=[]),
rowcount=1.7976931348623157E308, cumulative cost={inf}
rel#2117:DrillProjectRel.LOGICAL.ANY([]).[](child=rel#2095:Subset#1.LOGICAL.ANY([]).[],user_port=$1),
rowcount=20.0, cumulative cost={40.0 rows, 84.0 cpu, 0.0 io, 0.0 network}
[05104eb3-8320-47c0-b9d2-710840e44a29]
{code}
Here are 2 suggestions/asks:
1. Do we support non-equal join? if yes , we should fix this issue.
2. If we do not support, the error message should be more clear and simpler to
tell something like "non-equal join is not supported in current version...."
--
This message was sent by Atlassian JIRA
(v6.2#6252)
