mysql创建索引

经常发现建立索引后，sql的执行计划不会发生变化，对表运行下述语句后，才能起作用（在plsql 命令窗口中运行）

execute dbms_stats.gather_table_stats(ownname => 'USER_NAME',tabname => 'TABLE_NAME' ,estimate_percent => null ,method_opt => 'for all indexed columns' ,cascade => true ,degree => 1);


索引不起作用的例子

select b.logon_id,
                           count(logon_id) as shop_count,
                           max(b.add_date) as last_view_date
                      from shop_logon_view_pd_info b
                     group by b.logon_id;

                     select /*+ index(idx_lvpi_lid_adate)*/
 b.logon_id,
 count(logon_id) as shop_count,
 max(b.add_date) as last_view_date
  from shop_logon_view_pd_info b
 group by b.logon_id

-- 对于产品访问表shop_logon_view_pd_info做group by统计（记录数，最新日期），由于count这些方法会扫描所有数据行，因此虽然对表建立了logon_id和add_date的组合所以，在oracle CBO还是会自动选择走全表扫描，及时使用hint语法强制指定还是走全表扫描

表fsys_model_param一共有1229993条数据
查询:

select * from fsys_model_param where model_id = '1682'

消耗24.56秒

创建索引

create index idx_fsys_model_param_model_id on fsys_model_param(model_id);
drop index idx_fsys_model_param_model_id on fsys_model_param;

再次查询

select * from fsys_model_param where model_id = '1682'

消耗0.00秒,基本不消耗时间

查看建立的索引:

show index from fsys_model_info;