database/mariadb-columnstore-engine
1
0
Fork 0
mirror of https://github.com/mariadb-corporation/mariadb-columnstore-engine.git synced 2025-07-30 19:23:07 +03:00
Code Activity

MCOL-641 Refactored MultiplicationOverflowCheck but it still has flaws.

Browse Source 
Introduced fDecimalOverflowCheck to enable/disable overflow check.

Add support into a FunctionColumn.

Low level scanning crashes on medium sized data sets.
This commit is contained in:
Roman Nozdrin
2020-03-24 13:41:28 +00:00
parent 74b64eb4f1
commit b5534eb847
13 changed files with 528 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files

33
dbcon/execplan/functioncolumn.h
View File

@ -246,15 +246,34 @@ public:
{
IDB_Decimal decimal = fFunctor->getDecimalVal(row, fFunctionParms, isNull, fOperationType);
if (fResultType.scale == decimal.scale)
if (UNLIKELY(fResultType.colWidth == utils::MAXLEGACYWIDTH
&& fResultType.scale == decimal.scale))
return decimal;
if (fResultType.scale > decimal.scale)
decimal.value *= IDB_pow[fResultType.scale - decimal.scale];
else
decimal.value = (int64_t)(decimal.value > 0 ?
(double)decimal.value / IDB_pow[decimal.scale - fResultType.scale] + 0.5 :
(double)decimal.value / IDB_pow[decimal.scale - fResultType.scale] - 0.5);
if (LIKELY(fResultType.colWidth == datatypes::MAXDECIMALWIDTH))
{
decimal.s128Value =
(datatypes::Decimal::isWideDecimalType(decimal.precision)) ?
decimal.s128Value : decimal.value;
int128_t scaleMultiplier, result;
int32_t scaleDiff = fResultType.scale - decimal.scale;
datatypes::getScaleDivisor(scaleMultiplier, abs(scaleDiff));
// WIP MCOL-641 Unconditionall overflow check
datatypes::MultiplicationOverflowCheck mul;
decimal.s128Value = (scaleDiff > 0
&& mul(decimal.s128Value, scaleMultiplier, result))
? result : decimal.s128Value / scaleMultiplier;
}
else if (fResultType.colWidth == utils::MAXLEGACYWIDTH)
{
if (fResultType.scale > decimal.scale)
decimal.value *= IDB_pow[fResultType.scale - decimal.scale];
else
decimal.value = (int64_t)(decimal.value > 0 ?
(double)decimal.value / IDB_pow[decimal.scale - fResultType.scale] + 0.5 :
(double)decimal.value / IDB_pow[decimal.scale - fResultType.scale] - 0.5);
}
decimal.scale = fResultType.scale;
decimal.precision = fResultType.precision;