产品动态
产品公告
outline "sql" set outline_info "outline";outline reset ""; outline reset all;outline flush;create table t1(a int, b int, c int, primary key(a));create table t2(a int, b int, c int, unique key idx2(a));create table t3(a int, b int, c int, unique key idx3(a));
参数名 | 动态 | 类型 | 默认 | 参数值范围 | 说明 |
cdb_opt_outline_enabled | yes | bool | false | true/false | 是否打开 outline 功能 |
outline "sql" set outline_info "outline";，注意 outline_info 后的字符串应该以 "OUTLINE:" 开头，"OUTLINE:" 之后为加入 HINT 之后的 SQL。如给 select * from t1, t2 where t1.a = t2.a 这条 SQL 的 t2 表加上 a 列上的索引。outline "select * from t1, t2 where t1.a = t2.a" set outline_info "OUTLINE:select * from t1, t2 use index(idx2) where t1.a = t2.a";
outline "sql" set outline_info "outline";，注意 outline_info 后的字符串应该以 "OPT:" 开头，"OPT:" 之后为需要加入的 optimizer hint 信息。如给 select * from t1 where t1.a in (select b from t2) 这条 SQL 指定 MATERIALIZATION/DUPSWEEDOUT 的 SEMIJOIN。outline "select * from t1 where t1.a in (select b from t2)" set outline_info "OPT:2#qb_name(qb2)";outline "select * from t1 where t1.a in (select b from t2)" set outline_info "OPT:1#SEMIJOIN(@qb2 MATERIALIZATION, DUPSWEEDOUT)";
ie. "OPT:1#max_execution_time(1000)"）。outline "sql" set outline_info "outline";，注意 outline_info 后的字符串应该以 "INDEX:" 开头，"INDEX:" 之后为需要加入的 index hint 信息。
下面举个例子：给 select * from t1 where t1.a in (select t1.a from t1 where t1.b in (select t1.a from t1 left join t2 on t1.a = t2.a)) 这条 SQL 的 query block 3 上数据库 test 下 t1 表增加 USE INDEX 的索引 idx1，类型为 FOR JOIN。outline "select * from t1 where t1.a in (select t1.a from t1 where t1.b in (select t1.a from t1 left join t2 on t1.a = t2.a))" set outline_info "INDEX:3#test#t1#idx1#1#0";
ie. "INDEX:2#test#t2#idx2#1#1" 表示将第2个 query block 中的 test.t2 表中绑定类型为 USE INDEX FOR JOIN 的 idx2 索引）。outline reset "sql";，如将 select * from t1, t2 where t1.a = t2.a 的 outline 信息删除：outline reset "select * from t1, t2 where t1.a = t2.a";。outline reset all，执行语句为：outline reset all;。文档反馈