mysql分组汇总后的数据结果与预期出现差异

问题场景：

mysql8
Innodb
1.业务表A有88个字段，大致有1000w条数据；
2.对A表进行分组汇总统计数据,查询的结果与预期结果出现差异；

---

问题描述

forecast_month值有202305/202306/202307/202308
1.按照下面sql把分组后的数据插入B表;
2.比对数据
select forecast_month,sum(covered_by_supply_qty) from A group by forecast_month;
select forecast_month,sum(covered_by_supply_qty) from B group by forecast_month;
会出现两边结果有不一致的情况，但是总量是一样的。
如A表
202305   10000
202306   20000
202307   10000
202308   50000
B表
202305   10100   		比A表多100
202306   19900			比A表少100
202307   10000
202308   50000

insert into B (crd_season_year_cd, crd_month, forecast_month, contract_crd_season_year, contract_crd_month, contract_forecast_month, type_cd, division_cd, owner_group, big_group, plant_cd, prod_lifecycle, ord_type_grp_desc, block, delivery_block, credit_hold, launch_flag, eta_late, eta_overdue, ppk_flag, shippable, forecast_shippable, pass_cxl, account_tier, city_special_case, type_cd_flag, open_delivery_qty, open_delivery_amt_usd, ready_for_delivery_on_hand_inventory_qty, ready_for_delivery_on_hand_inventory_amt_usd, covered_by_supply_qty, covered_by_supply_amt_usd, no_supply_qty, no_supply_amt_usd, reporting_order_qty, reporting_order_amt_usd)
 select
            crd_season_year_cd,
            crd_month,
            forecast_month,
            contract_crd_season_year,
            contract_crd_month,
            contract_forecast_month,
            type_cd,
            division_cd,
            owner_group,
            big_group,
            plant_cd,
            prod_lifecycle,
            ord_type_grp_desc,
            block,
            delivery_block,
            credit_hold,
            launch_flag,
            eta_late,
            eta_overdue,
            ppk_flag,
            shippable,
            forecast_shippable,
            pass_cxl,
            account_tier,
            city_special_case,
            type_cd_flag,
            sum(open_delivery_qty) open_delivery_qty,
            sum(open_delivery_qty * wholesale_trans_usd) open_delivery_amt_usd,
            sum(ready_for_delivery_on_hand_inventory_qty) ready_for_delivery_on_hand_inventory_qty,
            sum(ready_for_delivery_on_hand_inventory_qty * wholesale_trans_usd) ready_for_delivery_on_hand_inventory_amt_usd,
            sum(covered_by_supply_qty) covered_by_supply_qty,
            sum(covered_by_supply_amt_usd) covered_by_supply_amt_usd,
            sum(no_supply_qty) no_supply_qty,
            sum(no_supply_amt_usd) no_supply_amt_usd,
            sum(reporting_order_qty) reporting_order_qty,
            sum(reporting_order_qty * wholesale_trans_usd) reporting_order_amt_usd
        from
            A str
        group by
            crd_season_year_cd,
            crd_month,
            forecast_month,
            contract_crd_season_year,
            contract_crd_month,
            contract_forecast_month,
            type_cd,
            division_cd,
            owner_group,
            big_group,
            plant_cd,
            prod_lifecycle,
            ord_type_grp_desc,
            block,
            delivery_block,
            credit_hold,
            launch_flag,
            eta_late,
            eta_overdue,
            ppk_flag,
            shippable,
            forecast_shippable,
            pass_cxl,
            account_tier,
            city_special_case,
            type_cd_flag

---

问题分析：

上面的sql去掉一个sum字段或者group by字段发现数据又对上了，但是随着原始数据发生变化数据对不上的情况又出现了，最后发现是因为group by的字段两两或三三之间都是相同的值，如

至于具体原因就需要看mysql源码去分析了，目前也不知道，知道的可以留言讨论，抱拳了。

---

解决方案：

解决也很简单，使用CONCAT函数给group by字段拼接一个字符串。

select
            crd_season_year_cd,
            crd_month,
            forecast_month,
            contract_crd_season_year,
            contract_crd_month,
            contract_forecast_month,
            type_cd,
            division_cd,
            owner_group,
            big_group,
            plant_cd,
            prod_lifecycle,
            ord_type_grp_desc,
            block,
            delivery_block,
            credit_hold,
            launch_flag,
            eta_late,
            eta_overdue,
            ppk_flag,
            shippable,
            forecast_shippable,
            pass_cxl,
            account_tier,
            city_special_case,
            type_cd_flag,
            sum(open_delivery_qty) open_delivery_qty,
            sum(open_delivery_qty * wholesale_trans_usd) open_delivery_amt_usd,
            sum(ready_for_delivery_on_hand_inventory_qty) ready_for_delivery_on_hand_inventory_qty,
            sum(ready_for_delivery_on_hand_inventory_qty * wholesale_trans_usd) ready_for_delivery_on_hand_inventory_amt_usd,
            sum(covered_by_supply_qty) covered_by_supply_qty,
            sum(covered_by_supply_amt_usd) covered_by_supply_amt_usd,
            sum(no_supply_qty) no_supply_qty,
            sum(no_supply_amt_usd) no_supply_amt_usd,
            sum(reporting_order_qty) reporting_order_qty,
            sum(reporting_order_qty * wholesale_trans_usd) reporting_order_amt_usd
        from
            A str
        group by
            crd_season_year_cd ,
            crd_month ,
            forecast_month ,
            CONCAT('4-',contract_crd_season_year) ,
            CONCAT('5-',contract_crd_month) ,
            CONCAT('6-',contract_forecast_month) ,
            type_cd ,
            division_cd ,
            owner_group ,
            big_group ,
            plant_cd ,
            prod_lifecycle ,
            ord_type_grp_desc ,
            CONCAT('14-',block),
            CONCAT('15-',delivery_block) ,
            CONCAT('16-',credit_hold) ,
            CONCAT('17-',launch_flag) ,
            CONCAT('18-',eta_late) ,
            CONCAT('19-',eta_overdue) ,
            CONCAT('20-',ppk_flag) ,
            CONCAT('21-',shippable) ,
            CONCAT('22-',forecast_shippable) ,
            CONCAT('23-',pass_cxl) ,
            CONCAT('24-',account_tier) ,
            CONCAT('25-',city_special_case) ,
            CONCAT('26-',type_cd_flag)```