hcq/oceanbase
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

[FEAT MERGE]4_1_sql_feature

Browse Source 
Co-authored-by: leslieyuchen <leslieyuchen@gmail.com>
Co-authored-by: Charles0429 <xiezhenjiang@gmail.com>
Co-authored-by: raywill <hustos@gmail.com>
This commit is contained in:
obdev
2023-01-28 16:01:26 +08:00
committed by ob-robot
parent 3080f2b66f
commit 2d19a9d8f5
846 changed files with 161957 additions and 116661 deletions
Download Patch File Download Diff File Expand all files Collapse all files

158
src/sql/engine/expr/ob_expr_regexp_context.h
View File

@ -19,11 +19,10 @@
#include <sys/types.h>
#include <assert.h>
#include "lib/charset/ob_charset.h"
#include "lib/regex/regex/regalone.h"
#include "lib/regex/regex/ob_regex.h"
#include <icu/i18n/unicode/uregex.h>
#include "sql/engine/expr/ob_expr_operator.h"
// this regex is compatible with mysql 5.6.16
// this regex is compatible with mysql 8.0
namespace oceanbase
{
@ -36,104 +35,97 @@ public:
ObExprRegexContext();
virtual ~ObExprRegexContext();
public:
static const char *icu_version_string() { return U_ICU_VERSION; }
inline bool is_inited() const { return inited_; }
void destroy();
void reset();
// The previous regex compile result can be used if pattern not change, if %reusable is true.
// %string_buf must be the same with previous init too if %reusable is true.
int init(const common::ObString &origin_pattern,
int cflags,
common::ObExprStringBuf &string_buf,
const bool reusable);
int init(ObExprStringBuf &string_buf,
ObSQLSessionInfo *session_info,
const ObString &origin_pattern,
const uint32_t cflags,
const bool reusable,
const ObCollationType cs_type);
int match(const common::ObString &text,
int64_t start_offset,
bool& is_match,
common::ObExprStringBuf &string_buf) const;
int substr(const common::ObString &text,
int64_t occurrence,
int64_t subexpr,
common::ObString &sub,
common::ObExprStringBuf &string_buf,
bool from_begin,
bool from_end,
common::ObSEArray<uint32_t, 4> &begin_locations) const;
int count_match_str(const common::ObString &text,
int64_t subexpr,
int64_t &sub,
common::ObExprStringBuf &string_buf,
bool from_begin,
bool from_end,
common::ObSEArray<uint32_t, 4> &begin_locations) const;
int instr(const common::ObString &text,
int64_t occurrence,
int64_t return_option,
int64_t subexpr,
int64_t &sub,
common::ObExprStringBuf &string_buf,
bool from_begin,
bool from_end,
common::ObSEArray<uint32_t, 4> &begin_locations) const;
int like(const common::ObString &text,
int64_t occurrence,
bool &sub,
common::ObExprStringBuf &string_buf,
bool from_begin,
bool from_end,
common::ObSEArray<uint32_t, 4> &begin_locations) const;
int replace_substr(const common::ObString &text_string,
const common::ObString &replace_string,
int64_t occurrence,
common::ObExprStringBuf &string_buf,
common::ObIArray<size_t> &ch_position,
common::ObString &sub,
bool from_begin,
bool from_end,
common::ObSEArray<uint32_t, 4> &begin_locations) const;
int replace(const common::ObString &text,
const common::ObString &to,
common::ObSEArray<uint32_t, 4> &locations_and_length,
common::ObExprStringBuf &string_buf,
common::ObIArray<common::ObSEArray<common::ObString, 8> > &subexpr_arrays,
common::ObString &sub) const;
int pre_process_replace_str(const common::ObString &text,
const common::ObString &to,
common::ObExprStringBuf &string_buf,
common::ObIArray<common::ObSEArray<common::ObString, 8> > &subexpr_arrays,
common::ObIArray<common::ObString> &subexpr_array) const;
int extract_subpre_string(const wchar_t *wc_text,
int64_t wc_length,
int64_t start_pos,
ob_regmatch_t pmatch[],
uint64_t pmatch_size,
common::ObExprStringBuf &string_buf,
common::ObIArray<common::ObString> &subexpr_array) const;
int match(ObExprStringBuf &string_buf,
const ObString &text,
const int64_t start,
bool &result) const;
int find(ObExprStringBuf &string_buf,
const ObString &text,
const int64_t start,
const int64_t occurrence,
const int64_t return_option,
const int64_t subexpr,
int64_t &result) const;
int count(ObExprStringBuf &string_buf,
const ObString &text,
const int32_t start,
int64_t &result) const;
int substr(ObExprStringBuf &string_buf,
const ObString &text,
const int64_t start,
const int64_t occurrence,
const int64_t subexpr,
ObString &result) const;
int replace(ObExprStringBuf &string_buf,
const ObString &text_string,
const ObString &replace_string,
const int64_t start,
const int64_t occurrence,
ObString &result) const;
int append_head(ObExprStringBuf &string_buf,
const int32_t current_pos,
UChar *&replace_buff,
int32_t &buff_size,
int32_t &buff_pos) const;
int append_replace_str(ObExprStringBuf &string_buf,
const UChar *u_replace,
const int32_t u_replace_length,
UChar *&replace_buff,
int32_t &buff_size,
int32_t &buff_pos) const;
int append_tail(ObExprStringBuf &string_buf,
UChar *&replace_buff,
int32_t &buff_size,
int32_t &buff_pos) const;
static int get_regexp_flags(const ObString &match_param,
const bool is_case_sensitive,
uint32_t &flags);
static int check_need_utf8(ObRawExpr *expr, bool &is_nstring);
TO_STRING_KV(K_(inited));
private:
void reset_reg();
int getwc(const common::ObString &text,
wchar_t *&wc,
int64_t &wc_length,
common::ObExprStringBuf &string_buf) const;
int w2c(const wchar_t *wc,
int64_t length,
char *&chr,
int64_t &chr_length,
common::ObExprStringBuf &string_buf) const;
int convert_reg_err_code_to_ob_err_code(int reg_err) const;
int preprocess_pattern(common::ObExprStringBuf &string_buf,
const common::ObString &origin_pattern,
common::ObString &pattern);
int check_icu_regexp_status(UErrorCode u_error_code, const UParseError *parse_error = NULL) const;
int get_valid_unicode_string(ObExprStringBuf &string_buf,
const ObString &origin_str,
UChar *&u_str,
int32_t &u_str_len) const;
int get_valid_replace_string(ObIAllocator &alloc,
const ObString &origin_replace,
UChar *&u_replace,
int32_t &u_replace_len) const;
private:
bool inited_;
ob_regex_t reg_;
ObInplaceAllocator pattern_allocator_;
common::ObString pattern_;
int cflags_;
ObInplaceAllocator pattern_wc_allocator_;
URegularExpression *regexp_engine_;
};
}
}