database/mariadb-columnstore-engine
1
0
Fork 0
mirror of https://github.com/mariadb-corporation/mariadb-columnstore-engine.git synced 2025-07-29 08:21:15 +03:00
Code Activity

MCOL-4589 Follow up on MCOL-4543 to optimize out non-referenced

Browse Source 
columns from a subquery involving UNIONs.
This commit is contained in:
Gagan Goel
2021-03-08 07:23:51 -05:00
parent 8a03e6c7d1
commit dfa9657f2e
3 changed files with 100 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files

63
dbcon/mysql/ha_from_sub.cpp
View File

@ -57,7 +57,7 @@ void derivedTableOptimization(THD* thd, SCSEP& csep)
for (uint i = 0; i < derivedTbList.size(); i++) for (uint i = 0; i < derivedTbList.size(); i++)
{ {
CalpontSelectExecutionPlan* plan = dynamic_cast<CalpontSelectExecutionPlan*>(derivedTbList[i].get()); CalpontSelectExecutionPlan* plan = reinterpret_cast<CalpontSelectExecutionPlan*>(derivedTbList[i].get());
CalpontSelectExecutionPlan::ReturnedColumnList cols = plan->returnedCols(); CalpontSelectExecutionPlan::ReturnedColumnList cols = plan->returnedCols();
vector<CalpontSelectExecutionPlan::ReturnedColumnList> unionColVec; vector<CalpontSelectExecutionPlan::ReturnedColumnList> unionColVec;
@ -73,7 +73,7 @@ void derivedTableOptimization(THD* thd, SCSEP& csep)
for (uint j = 0; j < plan->unionVec().size(); j++) for (uint j = 0; j < plan->unionVec().size(); j++)
{ {
unionColVec.push_back( unionColVec.push_back(
dynamic_cast<CalpontSelectExecutionPlan*>(plan->unionVec()[j].get())->returnedCols()); reinterpret_cast<CalpontSelectExecutionPlan*>(plan->unionVec()[j].get())->returnedCols());
} }
} }
@ -82,7 +82,7 @@ void derivedTableOptimization(THD* thd, SCSEP& csep)
for (uint j = 0; j < plan->unionVec().size(); j++) for (uint j = 0; j < plan->unionVec().size(); j++)
{ {
if (dynamic_cast<CalpontSelectExecutionPlan*>(plan->unionVec()[j].get())->tableList().empty()) if (reinterpret_cast<CalpontSelectExecutionPlan*>(plan->unionVec()[j].get())->tableList().empty())
{ {
horizontalOptimization = false; horizontalOptimization = false;
break; break;
@ -111,6 +111,7 @@ void derivedTableOptimization(THD* thd, SCSEP& csep)
// of the subquery in ExeMgr. // of the subquery in ExeMgr.
// This will be addressed in future. // This will be addressed in future.
CalpontSelectExecutionPlan::ReturnedColumnList nonConstCols; CalpontSelectExecutionPlan::ReturnedColumnList nonConstCols;
vector<CalpontSelectExecutionPlan::ReturnedColumnList> nonConstUnionColVec(unionColVec.size());
int64_t lastNonConstIndex = -1; int64_t lastNonConstIndex = -1;
@ -122,7 +123,7 @@ void derivedTableOptimization(THD* thd, SCSEP& csep)
if (cols[i]->derivedRefCol()) if (cols[i]->derivedRefCol())
cols[i]->derivedRefCol()->decRefCount(); cols[i]->derivedRefCol()->decRefCount();
if ((lastNonConstIndex == -1) && unionColVec.empty()) if (lastNonConstIndex == -1)
{ {
SimpleColumn* sc = dynamic_cast<SimpleColumn*>(cols[i].get()); SimpleColumn* sc = dynamic_cast<SimpleColumn*>(cols[i].get());
@ -134,13 +135,28 @@ void derivedTableOptimization(THD* thd, SCSEP& csep)
else else
{ {
cols[i].reset(new ConstantColumn(val)); cols[i].reset(new ConstantColumn(val));
(dynamic_cast<ConstantColumn*>(cols[i].get()))->timeZone(thd->variables.time_zone->get_name()->ptr()); (reinterpret_cast<ConstantColumn*>(cols[i].get()))->timeZone(thd->variables.time_zone->get_name()->ptr());
} }
for (uint j = 0; j < unionColVec.size(); j++) for (uint j = 0; j < unionColVec.size(); j++)
{ {
unionColVec[j][i].reset(new ConstantColumn(val)); if (lastNonConstIndex == -1)
(dynamic_cast<ConstantColumn*>(unionColVec[j][i].get()))->timeZone(thd->variables.time_zone->get_name()->ptr()); {
CalpontSelectExecutionPlan* unionSubPlan =
reinterpret_cast<CalpontSelectExecutionPlan*>(plan->unionVec()[j].get());
SimpleColumn* sc = dynamic_cast<SimpleColumn*>(unionSubPlan->returnedCols()[i].get());
if (sc && (unionSubPlan->columnMap().count(sc->columnName()) == 1))
{
unionSubPlan->columnMap().erase(sc->columnName());
}
}
else
{
unionColVec[j][i].reset(new ConstantColumn(val));
(reinterpret_cast<ConstantColumn*>(unionColVec[j][i].get()))->timeZone(thd->variables.time_zone->get_name()->ptr());
}
} }
} }
else if (lastNonConstIndex == -1) else if (lastNonConstIndex == -1)
@ -155,27 +171,36 @@ void derivedTableOptimization(THD* thd, SCSEP& csep)
if (!cols.empty()) if (!cols.empty())
{ {
cols[0].reset(new ConstantColumn(val)); cols[0].reset(new ConstantColumn(val));
(dynamic_cast<ConstantColumn*>(cols[0].get()))->timeZone(thd->variables.time_zone->get_name()->ptr()); (reinterpret_cast<ConstantColumn*>(cols[0].get()))->timeZone(thd->variables.time_zone->get_name()->ptr());
nonConstCols.push_back(cols[0]); nonConstCols.push_back(cols[0]);
for (uint j = 0; j < unionColVec.size(); j++)
{
unionColVec[j][0].reset(new ConstantColumn(val));
(reinterpret_cast<ConstantColumn*>(unionColVec[j][0].get()))->timeZone(thd->variables.time_zone->get_name()->ptr());
nonConstUnionColVec[j].push_back(unionColVec[j][0]);
}
} }
} }
else else
{ {
nonConstCols.assign(cols.begin(), cols.begin() + lastNonConstIndex + 1); nonConstCols.assign(cols.begin(), cols.begin() + lastNonConstIndex + 1);
for (uint j = 0; j < unionColVec.size(); j++)
{
nonConstUnionColVec[j].assign(unionColVec[j].begin(), unionColVec[j].begin() + lastNonConstIndex + 1);
}
} }
// set back // set back
if (unionColVec.empty()) plan->returnedCols(nonConstCols);
{
plan->returnedCols(nonConstCols);
}
else
{
plan->returnedCols(cols);
}
for (uint j = 0; j < unionColVec.size(); j++) for (uint j = 0; j < unionColVec.size(); j++)
dynamic_cast<CalpontSelectExecutionPlan*>(plan->unionVec()[j].get())->returnedCols(unionColVec[j]); {
CalpontSelectExecutionPlan* unionSubPlan =
reinterpret_cast<CalpontSelectExecutionPlan*>(plan->unionVec()[j].get());
unionSubPlan->returnedCols(nonConstUnionColVec[j]);
}
} }
} }
@ -210,7 +235,7 @@ void derivedTableOptimization(THD* thd, SCSEP& csep)
for (uint i = 0; i < derivedTbList.size(); i++) for (uint i = 0; i < derivedTbList.size(); i++)
{ {
CalpontSelectExecutionPlan* plan = dynamic_cast<CalpontSelectExecutionPlan*>(derivedTbList[i].get()); CalpontSelectExecutionPlan* plan = reinterpret_cast<CalpontSelectExecutionPlan*>(derivedTbList[i].get());
CalpontSelectExecutionPlan::ReturnedColumnList derivedColList = plan->returnedCols(); CalpontSelectExecutionPlan::ReturnedColumnList derivedColList = plan->returnedCols();
mapIt = derivedTbFilterMap.find(plan->derivedTbAlias()); mapIt = derivedTbFilterMap.find(plan->derivedTbAlias());
@ -240,7 +265,7 @@ void derivedTableOptimization(THD* thd, SCSEP& csep)
for (uint j = 0; j < plan->unionVec().size(); j++) for (uint j = 0; j < plan->unionVec().size(); j++)
{ {
CalpontSelectExecutionPlan* unionPlan = CalpontSelectExecutionPlan* unionPlan =
dynamic_cast<CalpontSelectExecutionPlan*>(plan->unionVec()[j].get()); reinterpret_cast<CalpontSelectExecutionPlan*>(plan->unionVec()[j].get());
CalpontSelectExecutionPlan::ReturnedColumnList unionColList = unionPlan->returnedCols(); CalpontSelectExecutionPlan::ReturnedColumnList unionColList = unionPlan->returnedCols();
ParseTree* mainFilterForUnion = new ParseTree(); ParseTree* mainFilterForUnion = new ParseTree();
mainFilterForUnion->copyTree(*(mapIt->second)); mainFilterForUnion->copyTree(*(mapIt->second));

51
mtr/basic/r/mcol-4543.result
View File

@ -176,10 +176,59 @@ COUNT(a)
SELECT COUNT(b) FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q; SELECT COUNT(b) FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q;
COUNT(b) COUNT(b)
10 10
SELECT COUNT(b), COUNT(a) FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q;
COUNT(b) COUNT(a)
10 10
SELECT COUNT(a) FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q GROUP BY b ORDER BY b; SELECT COUNT(a) FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q GROUP BY b ORDER BY b;
COUNT(a) COUNT(a)
4 4
6 6
SELECT q1.a FROM (SELECT * FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q1_1) q1 JOIN
(SELECT * FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q2_1) q2 ON q1.a=q2.a ORDER BY 1;
a
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
SELECT q1.a, q2.b FROM (SELECT * FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q1_1) q1 JOIN
(SELECT * FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q2_1) q2 ON q1.a=q2.a ORDER BY 2 desc, 1 asc;
a b
3 2
3 2
3 2
3 2
4 2
4 2
4 2
4 2
5 2
5 2
5 2
5 2
1 1
1 1
1 1
1 1
2 1
2 1
2 1
2 1
SELECT "123" FROM (SELECT * FROM t1) q GROUP BY b ORDER BY b; SELECT "123" FROM (SELECT * FROM t1) q GROUP BY b ORDER BY b;
123 123
123 123
@ -188,4 +237,4 @@ SELECT "123" FROM (SELECT * FROM t1) q GROUP BY b;
123 123
123 123
123 123
DROP DATABASE IF EXISTS mcol4543; DROP DATABASE mcol4543;

7
mtr/basic/t/mcol-4543.test
View File

@ -51,7 +51,12 @@ SELECT tab1.a, tab2.b FROM t1 tab1 JOIN (SELECT * FROM t1) tab2 ON tab1.a=tab2.a
# Test subquery columns referenced/not-referenced when subqueries contain unions # Test subquery columns referenced/not-referenced when subqueries contain unions
SELECT COUNT(a) FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q; SELECT COUNT(a) FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q;
SELECT COUNT(b) FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q; SELECT COUNT(b) FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q;
SELECT COUNT(b), COUNT(a) FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q;
SELECT COUNT(a) FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q GROUP BY b ORDER BY b; SELECT COUNT(a) FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q GROUP BY b ORDER BY b;
SELECT q1.a FROM (SELECT * FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q1_1) q1 JOIN
(SELECT * FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q2_1) q2 ON q1.a=q2.a ORDER BY 1;
SELECT q1.a, q2.b FROM (SELECT * FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q1_1) q1 JOIN
(SELECT * FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q2_1) q2 ON q1.a=q2.a ORDER BY 2 desc, 1 asc;
# Patch for MCOL-4543 also optimizes out an unnecessary BPS projection in PrimProc # Patch for MCOL-4543 also optimizes out an unnecessary BPS projection in PrimProc
# that was happening earlier. The following 2 queries trigger this optimization. # that was happening earlier. The following 2 queries trigger this optimization.
@ -59,4 +64,4 @@ SELECT COUNT(a) FROM (SELECT * FROM t1 UNION ALL SELECT * FROM t1) q GROUP BY b
SELECT "123" FROM (SELECT * FROM t1) q GROUP BY b ORDER BY b; SELECT "123" FROM (SELECT * FROM t1) q GROUP BY b ORDER BY b;
SELECT "123" FROM (SELECT * FROM t1) q GROUP BY b; SELECT "123" FROM (SELECT * FROM t1) q GROUP BY b;
DROP DATABASE IF EXISTS mcol4543; DROP DATABASE mcol4543;