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-4188 Regression fixes for MCOL-641.

Browse Source 
1. Add wide decimal support to AggregateColumn::evaluate
and TreeNode::getDecimalVal().
2. Use the pm aggregate attributes to determine um aggregate
attributes in TupleAggregateStep::prep2PhasesAggregate.
This commit is contained in:
Gagan Goel
2020-11-19 18:56:40 -05:00
parent aa64d81cde
commit a159f8a0b6
8 changed files with 125 additions and 44 deletions
Download Patch File Download Diff File Expand all files Collapse all files

77
dbcon/execplan/treenode.h
View File

@ -1003,41 +1003,82 @@ inline IDB_Decimal TreeNode::getDecimalVal()
throw logging::InvalidConversionExcept("TreeNode::getDecimalVal: non-support conversion from binary string");
case CalpontSystemCatalog::BIGINT:
{
idbassert(fResultType.scale == 0);
if (fResultType.isWideDecimalPrecision())
fResult.decimalVal = IDB_Decimal(0, 0, fResultType.precision,
(int128_t)fResult.intVal);
else
fResult.decimalVal = IDB_Decimal(fResult.intVal, 0,
fResultType.precision);
break;
}
case CalpontSystemCatalog::MEDINT:
case CalpontSystemCatalog::INT:
case CalpontSystemCatalog::SMALLINT:
case CalpontSystemCatalog::TINYINT:
fResult.decimalVal.value = (int64_t)(fResult.intVal * pow((double)10.0, fResultType.scale));
fResult.decimalVal.scale = fResultType.scale;
fResult.decimalVal.precision = fResultType.precision;
{
idbassert(fResultType.scale == 0);
fResult.decimalVal = IDB_Decimal(fResult.intVal, 0,
fResultType.precision);
break;
}
case CalpontSystemCatalog::UBIGINT:
{
idbassert(fResultType.scale == 0);
if (fResultType.isWideDecimalPrecision())
fResult.decimalVal = IDB_Decimal(0, 0, fResultType.precision,
(int128_t)fResult.uintVal);
else
fResult.decimalVal = IDB_Decimal((int64_t)fResult.uintVal, 0,
fResultType.precision);
break;
}
case CalpontSystemCatalog::UMEDINT:
case CalpontSystemCatalog::UINT:
case CalpontSystemCatalog::USMALLINT:
case CalpontSystemCatalog::UTINYINT:
fResult.decimalVal.value = (int64_t)(fResult.uintVal * pow((double)10.0, fResultType.scale));
fResult.decimalVal.scale = fResultType.scale;
fResult.decimalVal.precision = fResultType.precision;
{
idbassert(fResultType.scale == 0);
fResult.decimalVal = IDB_Decimal((int64_t)fResult.uintVal, 0,
fResultType.precision);
break;
}
case CalpontSystemCatalog::LONGDOUBLE:
{
long double dlScaled = fResult.longDoubleVal;
if (fResultType.scale > 0)
{
long double dlScaled = fResult.longDoubleVal;
if (fResultType.scale > 0)
{
dlScaled= fResult.longDoubleVal * pow((double)10.0, fResultType.scale);
}
if (dlScaled > (double)MAX_BIGINT)
{
throw logging::InvalidConversionExcept("TreeNode::getDecimalVal: decimal overflow.");
}
fResult.decimalVal.value = (int64_t)roundl((fResult.longDoubleVal * pow((double)10.0, fResultType.scale)));
fResult.decimalVal.scale = fResultType.scale;
fResult.decimalVal.precision = fResultType.precision;
dlScaled = fResult.longDoubleVal * pow((double)10.0, fResultType.scale);
}
if ((dlScaled > (long double)INT64_MAX) || (dlScaled < (long double)(INT64_MIN)))
{
datatypes::TFloat128 temp((__float128)dlScaled);
fResult.decimalVal = IDB_Decimal(0, fResultType.scale,
fResultType.precision, static_cast<int128_t>(temp));
}
else
{
fResult.decimalVal = IDB_Decimal((int64_t)lroundl(dlScaled),
fResultType.scale, fResultType.precision);
}
break;
}
case CalpontSystemCatalog::DATE:
throw logging::InvalidConversionExcept("TreeNode::getDecimalVal: Invalid conversion from date.");