grafana数据过滤器不支持变量

我也不知道该记录在哪儿，事实上，这是一个grafana使用mysql查询数据的问题；

我的mysql是8.0.21版本；

我在grafana上设置两个变量，line_name和tower_name；然后使用sql语句去查询数据，但是有时候，tower_name变量是空的，而且tower_name对应的表的数据量比较大；这直接导致查询语句变成了这样：

# 性能问题还是比较严重
SELECT * FROM (SELECT
  obj_code as `pms编码`,
  obj_order as `名称`,
  p_name as `站线`,
  device_state as `状态`,
  city_name as `地市`,
  maintain_unitname as `运维分部`,
  cast(last_update_time as char(20)) as '更新时间',
  obj_id  as 'psrId资源id',
  pms_id as 'astId资产id',
  obj_type as '台账类型'
FROM
  r_device_tower
WHERE
  p_line in (
    SELECT
      obj_id
    FROM
      r_device_line
    WHERE
      line_name like '%某某线路%'
  ))as sub
  WHERE sub.名称 like '%%'

直接变成慢查询，其实原因出现在【sub.名称 like '%%'】，从结果上来说，要不要这个条件的结果都是一样的；但是mysql那边好像没有优化这样的查询。

然后grafana里面的【filter data by values】过滤器也是垃圾，不支持使用变量来进行过滤；

解决办法是：使用正则表达式：

# 性能问题还是比较严重
SELECT * FROM (SELECT
  obj_code as `pms编码`,
  obj_order as `名称`,
  p_name as `站线`,
  device_state as `状态`,
  city_name as `地市`,
  maintain_unitname as `运维分部`,
  cast(last_update_time as char(20)) as '更新时间',
  obj_id  as 'psrId资源id',
  pms_id as 'astId资产id',
  obj_type as '台账类型'
FROM
  r_device_tower
WHERE
  p_line in (
    SELECT
      obj_id
    FROM
      r_device_line
    WHERE
      line_name like '%$line_name%'
  ))as sub
  WHERE sub.名称 REGEX '.*$tower_name.*';

性能直线上升！