1
0
mirror of https://github.com/mariadb-corporation/mariadb-columnstore-engine.git synced 2025-04-18 21:44:02 +03:00

move convert to datatypes::Charset class

This commit is contained in:
Leonid Fedorov 2024-03-13 12:34:06 +00:00
parent c024bb0be5
commit 38c9b51a13
4 changed files with 60 additions and 61 deletions

View File

@ -543,97 +543,97 @@ INSERT INTO t2 values('こんにちは');
INSERT INTO t2 values('привет');
INSERT INTO t2 values('Γεια');
INSERT INTO t2 values('სალამი');
SELECT hello, hello regexp 'ん.ち' from t2;
SELECT hello, hello regexp 'ん.ち' FROM t2;
hello hello regexp 'ん.ち'
こんにちは 0
привет 0
Γεια 0
სალამი 0
SELECT hello, hello regexp 'и.е' from t2;
SELECT hello, hello regexp 'и.е' FROM t2;
hello hello regexp 'и.е'
こんにちは 0
привет 0
Γεια 0
სალამი 0
SELECT hello, hello regexp 'ε.α' from t2;
SELECT hello, hello regexp 'ε.α' FROM t2;
hello hello regexp 'ε.α'
こんにちは 0
привет 0
Γεια 0
სალამი 0
SELECT hello, hello regexp 'ა.ა' from t2;
SELECT hello, hello regexp 'ა.ა' FROM t2;
hello hello regexp 'ა.ა'
こんにちは 0
привет 0
Γεια 0
სალამი 0
SELECT hello, regexp_substr(hello, 'ん.ち') from t2;
SELECT hello, regexp_substr(hello, 'ん.ち') FROM t2;
hello regexp_substr(hello, 'ん.ち')
こんにちは
привет
Γεια
სალამი
SELECT hello, regexp_substr(hello, 'и.е') from t2;
SELECT hello, regexp_substr(hello, 'и.е') FROM t2;
hello regexp_substr(hello, 'и.е')
こんにちは
привет
Γεια
სალამი
SELECT hello, regexp_substr(hello, 'ε.α') from t2;
SELECT hello, regexp_substr(hello, 'ε.α') FROM t2;
hello regexp_substr(hello, 'ε.α')
こんにちは
привет
Γεια
სალამი
SELECT hello, regexp_substr(hello, 'ა.ა') from t2;
SELECT hello, regexp_substr(hello, 'ა.ა') FROM t2;
hello regexp_substr(hello, 'ა.ა')
こんにちは
привет
Γεια
სალამი
SELECT hello, regexp_instr(hello, 'ん.ち') from t2;
SELECT hello, regexp_instr(hello, 'ん.ち') FROM t2;
hello regexp_instr(hello, 'ん.ち')
こんにちは 0
привет 0
Γεια 0
სალამი 0
SELECT hello, regexp_instr(hello, 'и.е') from t2;
SELECT hello, regexp_instr(hello, 'и.е') FROM t2;
hello regexp_instr(hello, 'и.е')
こんにちは 0
привет 0
Γεια 0
სალამი 0
SELECT hello, regexp_instr(hello, 'ε.α') from t2;
SELECT hello, regexp_instr(hello, 'ε.α') FROM t2;
hello regexp_instr(hello, 'ε.α')
こんにちは 0
привет 0
Γεια 0
სალამი 0
SELECT hello, regexp_instr(hello, 'ა.ა') from t2;
SELECT hello, regexp_instr(hello, 'ა.ა') FROM t2;
hello regexp_instr(hello, 'ა.ა')
こんにちは 0
привет 0
Γεια 0
სალამი 0
SELECT hello, regexp_replace(hello, 'ん.ち', 'Достоевский') from t2;
SELECT hello, regexp_replace(hello, 'ん.ち', 'Достоевский') FROM t2;
hello regexp_replace(hello, 'ん.ち', 'Достоевский')
こんにちは こんにちは
привет привет
Γεια Γεια
სალამი სალამი
SELECT hello, regexp_replace(hello, 'и.е', 'Достоевский') from t2;
SELECT hello, regexp_replace(hello, 'и.е', 'Достоевский') FROM t2;
hello regexp_replace(hello, 'и.е', 'Достоевский')
こんにちは こんにちは
привет привет
Γεια Γεια
სალამი სალამი
SELECT hello, regexp_replace(hello, 'ε.α', 'Достоевский') from t2;
SELECT hello, regexp_replace(hello, 'ε.α', 'Достоевский') FROM t2;
hello regexp_replace(hello, 'ε.α', 'Достоевский')
こんにちは こんにちは
привет привет
Γεια Γεια
სალამი სალამი
SELECT hello, regexp_replace(hello, 'ა.ა', 'Достоевский') from t2;
SELECT hello, regexp_replace(hello, 'ა.ა', 'Достоевский') FROM t2;
hello regexp_replace(hello, 'ა.ა', 'Достоевский')
こんにちは こんにちは
привет привет

View File

@ -152,32 +152,31 @@ SELECT t1_TIME, REGEXP_REPLACE(t1_TIME, '22$', 'KittyCat') FROM t1 ORDER BY 1;
SET character_set_connection = 'utf8';
CREATE TABLE t2 (hello text) engine columnstore;
INSERT INTO t2 values('こんにちは');
INSERT INTO t2 values('привет');
INSERT INTO t2 values('Γεια');
INSERT INTO t2 values('სალამი');
SELECT hello, hello regexp 'ん.ち' from t2;
SELECT hello, hello regexp 'и.е' from t2;
SELECT hello, hello regexp 'ε.α' from t2;
SELECT hello, hello regexp 'ა.ა' from t2;
SELECT hello, hello regexp 'ん.ち' FROM t2;
SELECT hello, hello regexp 'и.е' FROM t2;
SELECT hello, hello regexp 'ε.α' FROM t2;
SELECT hello, hello regexp 'ა.ა' FROM t2;
SELECT hello, regexp_substr(hello, 'ん.ち') from t2;
SELECT hello, regexp_substr(hello, 'и.е') from t2;
SELECT hello, regexp_substr(hello, 'ε.α') from t2;
SELECT hello, regexp_substr(hello, 'ა.ა') from t2;
SELECT hello, regexp_substr(hello, 'ん.ち') FROM t2;
SELECT hello, regexp_substr(hello, 'и.е') FROM t2;
SELECT hello, regexp_substr(hello, 'ε.α') FROM t2;
SELECT hello, regexp_substr(hello, 'ა.ა') FROM t2;
SELECT hello, regexp_instr(hello, 'ん.ち') from t2;
SELECT hello, regexp_instr(hello, 'и.е') from t2;
SELECT hello, regexp_instr(hello, 'ε.α') from t2;
SELECT hello, regexp_instr(hello, 'ა.ა') from t2;
SELECT hello, regexp_instr(hello, 'ん.ち') FROM t2;
SELECT hello, regexp_instr(hello, 'и.е') FROM t2;
SELECT hello, regexp_instr(hello, 'ε.α') FROM t2;
SELECT hello, regexp_instr(hello, 'ა.ა') FROM t2;
SELECT hello, regexp_replace(hello, 'ん.ち', 'Достоевский') from t2;
SELECT hello, regexp_replace(hello, 'и.е', 'Достоевский') from t2;
SELECT hello, regexp_replace(hello, 'ε.α', 'Достоевский') from t2;
SELECT hello, regexp_replace(hello, 'ა.ა', 'Достоевский') from t2;
SELECT hello, regexp_replace(hello, 'ん.ち', 'Достоевский') FROM t2;
SELECT hello, regexp_replace(hello, 'и.е', 'Достоевский') FROM t2;
SELECT hello, regexp_replace(hello, 'ε.α', 'Достоевский') FROM t2;
SELECT hello, regexp_replace(hello, 'ა.ა', 'Достоевский') FROM t2;
# Clean UP
DROP DATABASE mcs228_db;

View File

@ -146,6 +146,23 @@ class Charset
Charset(CHARSET_INFO* cs = nullptr) : mCharset(cs ? cs : &my_charset_bin)
{
}
bool operator==(const Charset& rhs)
{
return rhs.getCharset().cs_name.str == getCharset().cs_name.str;
}
std::string convert(const std::string& from, const datatypes::Charset& fromCs) const
{
std::string result;
uint dummy_errors;
result.resize(from.size() * getCharset().mbmaxlen);
size_t resultingSize = my_convert(const_cast<char*>(result.c_str()), result.size(), &getCharset(), from.c_str(),
from.size(), &fromCs.getCharset(), &dummy_errors);
result.resize(resultingSize);
return result;
}
Charset(uint32_t charsetNumber);
void setCharset(uint32_t charsetNumber);
CHARSET_INFO& getCharset() const

View File

@ -39,42 +39,27 @@ using namespace execplan;
#include "errorcodes.h"
#include "idberrorinfo.h"
#include "errorids.h"
using namespace logging;
namespace
{
std::string csConvert(const std::string& from, CHARSET_INFO* to_cs, CHARSET_INFO* from_cs)
{
std::string result;
uint dummy_errors;
result.resize(from.size() * to_cs->mbmaxlen);
size_t resultingSize = my_convert(const_cast<char*>(result.c_str()), result.size(), to_cs, from.c_str(),
from.size(), from_cs, &dummy_errors);
result.resize(resultingSize);
return result;
}
using jp = jpcre2::select<char>;
struct PCREOptions
{
PCREOptions(execplan::CalpontSystemCatalog::ColType& ct);
CHARSET_INFO* dataCharset = &my_charset_utf8mb3_general_ci;
CHARSET_INFO* libraryCharset = &my_charset_utf8mb3_general_ci;
datatypes::Charset dataCharset = my_charset_utf8mb3_general_ci;
datatypes::Charset libraryCharset = my_charset_utf8mb3_general_ci;
jpcre2::Uint flags = 0;
bool conversionIsNeeded = false;
};
inline bool areSameCharsets(CHARSET_INFO* cs1, CHARSET_INFO* cs2)
{
return (cs1->cs_name.str == cs2->cs_name.str);
}
PCREOptions::PCREOptions(execplan::CalpontSystemCatalog::ColType& ct)
{
CHARSET_INFO* cs = ct.getCharset();
datatypes::Charset cs = ct.getCharset();
// TODO use system variable instead if hardcode default_regex_flags_pcre(_current_thd());
// PCRE2_DOTALL | PCRE2_DUPNAMES | PCRE2_EXTENDED | PCRE2_EXTENDED_MORE | PCRE2_MULTILINE | PCRE2_UNGREEDY;
@ -82,12 +67,11 @@ PCREOptions::PCREOptions(execplan::CalpontSystemCatalog::ColType& ct)
jpcre2::Uint defaultFlags = 0;
flags = (cs != &my_charset_bin ? (PCRE2_UTF | PCRE2_UCP) : 0) |
((cs->state & (MY_CS_BINSORT | MY_CS_CSSORT)) ? 0 : PCRE2_CASELESS) | defaultFlags;
((cs.getCharset().state & (MY_CS_BINSORT | MY_CS_CSSORT)) ? 0 : PCRE2_CASELESS) | defaultFlags;
// Convert text data to utf-8.
dataCharset = cs;
libraryCharset = cs == &my_charset_bin ? &my_charset_bin : &my_charset_utf8mb3_general_ci;
conversionIsNeeded = (cs != &my_charset_bin) && !areSameCharsets(cs, libraryCharset);
libraryCharset = cs == my_charset_bin ? my_charset_bin : my_charset_utf8mb3_general_ci;
}
struct RegExpParams
@ -99,8 +83,8 @@ struct RegExpParams
if (options.conversionIsNeeded)
return *this;
expression = csConvert(expression, options.libraryCharset, options.dataCharset);
pattern = csConvert(pattern, options.libraryCharset, options.dataCharset);
expression = options.libraryCharset.convert(expression, options.dataCharset);
pattern = options.libraryCharset.convert(pattern, options.dataCharset);
return *this;
}
};
@ -329,11 +313,10 @@ std::string Func_regexp_replace::getStrVal(rowgroup::Row& row, FunctionParm& fp,
const auto& replaceWithStr = replaceWith.unsafeStringRef();
if (options.conversionIsNeeded)
{
const auto& convertedReplaceToken = csConvert(replaceWithStr, options.libraryCharset, options.dataCharset);
const auto& convertedReplaceToken = options.libraryCharset.convert(replaceWithStr, options.dataCharset);
return re.replace(param.expression, convertedReplaceToken, "g");
}
return re.replace(param.expression, replaceWithStr, "g");
}
@ -391,8 +374,8 @@ std::string Func_regexp_instr::getStrVal(rowgroup::Row& row, FunctionParm& fp, b
return "0";
size_t offset = vec_soff[0];
size_t charNumber =
options.libraryCharset->numchars(param.expression.c_str(), param.expression.c_str() + offset);
size_t charNumber = options.libraryCharset.getCharset().numchars(param.expression.c_str(),
param.expression.c_str() + offset);
return std::to_string(charNumber + 1);
}