用户定义的分析器

示例

以下用 C 语言编写的函数将字母和数字字符处理为标记，并忽略除下划线 (_) 之外的所有特殊字符：

/*ARGSUSED*/
UDREXPORT
mi_lvarchar* tokenize_alnum(
    mi_lvarchar*    string,
    MI_FPARAM*      fparam)
{
    mi_integer      status = MI_OK;
    mi_lvarchar*    rtn = NULL;
    gl_mchar_t*     src = NULL;
    gl_mchar_t*     tgt = NULL;
    mi_integer      token = 0;
    gl_mchar_t*     s;
    gl_mchar_t*     r;

    ifx_gl_init();
    if (((src = (gl_mchar_t*)mi_lvarchar_to_string(string)) == NULL) ||
        ((tgt = (gl_mchar_t*)mi_alloc((strlen(src)*4)+1)) == NULL)) {
        status = MI_ERROR;
        goto cleanup;
    }
    s = src;
    r = tgt;
    while ((s != NULL) && (*s != '\0')) {
        if ((ifx_gl_ismalnum(s, IFX_GL_NO_LIMIT)) || (*s == '_')) {
            if (!token) {
                if (r != tgt) *r++ = ' ';
                *r++ = '[';
                token = 1;
            }
            ifx_gl_mbsncpy(r, s, IFX_GL_NULL, 1);
            r = ifx_gl_mbsnext(r, IFX_GL_NO_LIMIT);
        }
        else {
            if (token) {
                *r++ = ']';
                token = 0;
            }
        }
        s = ifx_gl_mbsnext(s, IFX_GL_NO_LIMIT);
    }
    if (token) *r++ = ']';
    *r = '\0';
    if ((rtn = mi_string_to_lvarchar((char*)tgt)) == NULL) {
        status = MI_ERROR;
        goto cleanup;
    }
cleanup:
    if ((status != MI_OK) &&
        (rtn != NULL)) {
        mi_var_free(rtn);
        rtn = NULL;
    }
    if (tgt != NULL) mi_free(tgt);
    if (src != NULL) mi_free(src);
    if (rtn == NULL) mi_fp_setreturnisnull(fparam, 0, MI_TRUE);
    return rtn;
}

以下语句注册函数，以使数据库服务器可以使用此函数：

CREATE FUNCTION tokenize_alnum (lvarchar)
    RETURNS lvarchar
    WITH (NOT VARIANT)
    EXTERNAL NAME "$GBASEDBTDIR/extend/myblade/myblade.bld(tokenize_alnum)" 
        LANGUAGE C;

使用 analyzer="udr.tokenize_alnum" 选项创建索引时，以下示例显示没有对除下划线之外的特殊字符建立索引：

quick! #$%&^^$## Brown fox under_score 
[quick] [Brown] [fox] [under_score]