firebird/firebird-mirror
firebird/firebird-mirror
8
0
Fork 1
mirror of https://github.com/FirebirdSQL/firebird.git synced 2025-01-23 14:03:07 +01:00
Code Issues

Improvement #6872 - Indexed STARTING WITH execution is very slow with UNICODE collation.

Browse Source
This commit is contained in:
Adriano dos Santos Fernandes 2021-06-25 15:37:25 -03:00
parent b98df5e0bc
commit de49ddbca0
2 changed files with 44 additions and 27 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

61
src/common/unicode_util.cpp
View File

@ -1336,10 +1336,6 @@ UnicodeUtil::Utf16Collation* UnicodeUtil::Utf16Collation::create(
icu->ucolSetAttribute(compareCollator, UCOL_STRENGTH, UCOL_SECONDARY, &status);
}
USet* contractions = icu->usetOpen(0, 0);
// status not verified here.
icu->ucolGetContractions(partialCollator, contractions, &status);
Utf16Collation* obj = FB_NEW Utf16Collation();
obj->icu = icu;
obj->tt = tt;
@ -1347,9 +1343,39 @@ UnicodeUtil::Utf16Collation* UnicodeUtil::Utf16Collation::create(
obj->compareCollator = compareCollator;
obj->partialCollator = partialCollator;
obj->sortCollator = sortCollator;
obj->contractions = contractions;
obj->contractionsCount = icu->usetGetItemCount(contractions);
obj->numericSort = isNumericSort;
obj->maxContractionsPrefixLength = 0;
USet* contractions = icu->usetOpen(1, 0);
// status not verified here.
icu->ucolGetContractions(partialCollator, contractions, &status);
int contractionsCount = icu->usetGetItemCount(contractions);
for (int contractionIndex = 0; contractionIndex < contractionsCount; ++contractionIndex)
{
UChar str[10];
UChar32 start, end;
status = U_ZERO_ERROR;
int len = icu->usetGetItem(contractions, contractionIndex, &start, &end, str, sizeof(str), &status);
if (len >= 2)
{
obj->maxContractionsPrefixLength = len - 1 > obj->maxContractionsPrefixLength ?
len - 1 : obj->maxContractionsPrefixLength;
for (int currentLen = 1; currentLen < len; ++currentLen)
{
string s(reinterpret_cast<const char*>(str), currentLen * 2);
if (!obj->contractionsPrefix.exist(s))
obj->contractionsPrefix.push(s);
}
}
}
icu->usetClose(contractions);
return obj;
}
@ -1357,8 +1383,6 @@ UnicodeUtil::Utf16Collation* UnicodeUtil::Utf16Collation::create(
UnicodeUtil::Utf16Collation::~Utf16Collation()
{
icu->usetClose(contractions);
icu->ucolClose(compareCollator);
icu->ucolClose(partialCollator);
icu->ucolClose(sortCollator);
@ -1409,28 +1433,16 @@ USHORT UnicodeUtil::Utf16Collation::stringToKey(USHORT srcLen, const USHORT* src
switch (key_type)
{
case INTL_KEY_PARTIAL:
{
coll = partialCollator;
// Remove last bytes of key if they are start of a contraction
// to correctly find in the index.
ConversionICU& cIcu(getConversionICU());
for (int i = 0; i < contractionsCount; ++i)
{
UChar str[10];
UErrorCode status = U_ZERO_ERROR;
int len = icu->usetGetItem(contractions, i, NULL, NULL, str, sizeof(str), &status);
if (len > SLONG(srcLenLong))
len = srcLenLong;
else
--len;
// safe cast - alignment not changed
if (cIcu.u_strCompare(str, len,
reinterpret_cast<const UChar*>(src) + srcLenLong - len, len, true) == 0)
for (int i = MIN(maxContractionsPrefixLength, srcLenLong); i > 0; --i)
{
srcLenLong -= len;
if (contractionsPrefix.exist(string(reinterpret_cast<const char*>(src + srcLenLong - i), i * 2)))
{
srcLenLong -= i;
break;
}
}
@ -1451,7 +1463,6 @@ USHORT UnicodeUtil::Utf16Collation::stringToKey(USHORT srcLen, const USHORT* src
}
break;
}
case INTL_KEY_UNIQUE:
coll = compareCollator;

10
src/common/unicode_util.h
View File

@ -31,6 +31,7 @@
#include "../common/IntlUtil.h"
#include "../common/os/mod_loader.h"
#include "../common/classes/fb_string.h"
#include "../common/classes/objects_array.h"
#undef U_SHOW_CPLUSPLUS_API
#define U_SHOW_CPLUSPLUS_API 0
@ -168,6 +169,11 @@ public:
Firebird::IntlUtil::SpecificAttributesMap& specificAttributes,
const Firebird::string& configInfo);
Utf16Collation()
: contractionsPrefix(*getDefaultMemoryPool())
{
}
~Utf16Collation();
USHORT keyLength(USHORT len) const;
@ -190,8 +196,8 @@ public:
UCollator* compareCollator;
UCollator* partialCollator;
UCollator* sortCollator;
USet* contractions;
int contractionsCount;
Firebird::SortedObjectsArray<Firebird::string> contractionsPrefix; // UTF-16 string
unsigned maxContractionsPrefixLength; // number of characters
bool numericSort;
};