得到的数据包大于'max_allowed_pa​​cket'字节

Question

我有一个查询为大约5000个select语句执行UNION ALL。当我运行这个我得到这个错误

Got a packet bigger than 'max_allowed_packet' bytes

基于类似的其他问题，我更改了my.ini文件中的值并重新启动了服务器。我已将值设置为1GB，这是mysql接受的最大值。我可以看到服务器拾取了1GB的新值。我可以通过运行它来检查它：

SHOW VARIABLES LIKE 'max_allowed_packet';

我仍然继续收到此错误。我将实际的SQL语句保存到文本文件中，只保存了4.4mb。

我该怎么做才能解决这个问题。

示例选择声明：其中大约有5000个。唯一改变的是id

SELECT 
  SUM(postpay_GA + prepay_ga) AS GA,
  CAST(SUM(EDGE_NUM) / SUM(EDGE_DEN) AS DECIMAL (10, 4)) * 100 AS EDGE,
  SUM(ACC_REV) / SUM(ATR_D) AS ARD,
  (CAST(SUM(promot) AS DECIMAL (10, 4)) / SUM(surv) - CAST(SUM(DETRAC) AS DECIMAL (10, 4)) / SUM(surv)) * 100 AS NPS,
  SUM(Renew) AS GR 
FROM
  store_dashboard_data_dly 
WHERE ACTIVITY_MTH BETWEEN '2014-09-01' AND '2014-09-30' 
  AND region_cd = **'NE'** 
  AND logic = 'eis' 
UNION
ALL 
SELECT 
  SUM(quota * (metric_id = 1)),
  AVG(NULLIF(quota * (metric_id = 7), 0)),
  AVG(NULLIF(quota * (metric_id = 10), 0)),
  AVG(NULLIF(quota * (metric_id = 14), 0)),
  SUM(quota * (metric_id = 2)) 
FROM
  rop_budget 
WHERE id = **'NE'** 
  AND start_date BETWEEN '2014-09-01' AND '2014-09-30' 

Answer 1

您可以使用group by

以更简单的方式获得所有结果

select
    x.ga,
    x.edge,
    x.ard,
    x.nps,
    x.gr
from (
    select
        d.region_cd,
        sum(d.postpay_ga + d.prepay_ga) as ga,
        cast(sum(d.edge_num) / sum(d.edge_den) as decimal(10, 4)) * 100 as edge,
        sum(d.acc_rev) / sum(d.atr_d) as ard,
        (cast(sum(d.promot) as decimal(10, 4)) / sum(d.surv) 
          - cast(sum(d.detrac) as decimal(10, 4)) / sum(d.surv)) * 100 as nps,
        sum(d.Renew) as gr,
        1 as ord
    from
        store_dashboard_data_dly as d
    where
        d.activity_mth between '2014-09-01' and '2014-09-30' and
        d.region_cd in ('NE', 'R2', 'R3', ...) and
        d.logic = 'eis'
    group by
        d.region_cd
    union all
    select
        id,
        sum(quota * (metric_id = 1)),
        avg(nullif(quota * (metric_id = 7), 0)),
        avg(nullif(quota * (metric_id = 10), 0)),
        avg(nullif(quota * (metric_id = 14), 0)),
        sum(quota * (metric_id = 2)),
        2
    from
        rop_budget 
    where
        id in ('NE', 'R2', 'R3', ...) and
        start_date between '2014-09-01' and '2014-09-30'
) x
order by
    x.id,
    x.ord

如果您希望按照定义查询的相同顺序排序结果，则可以使用region_cd, sequence_no创建临时表。插入区域/序列，然后使用它连接报告表。