使用过程根据列中的状态计算大数据

Question

table 中有 770 万条记录（并且增长迅速），统计所有记录最多需要 19 秒。

我的要求是在 1 秒内加载它，请帮助我....

我的 table 结构是 ->

id  |campaign_id  | disposition |   action    create_date | more 16 columns...
1   |     3       |   CANCEL    |     NULL    2020-08-12  |
2   |     5       |   ANSWER    |     DNC     2020-08-13  |
3   |     3       |   ANSWER    |     SIP     2020-08-14  |

我的程序是->

proc_campaign_report(campaignId int(10), reportTable varchar(20))
BEGIN
 SET @SELECTFROM = CONCAT("
        select  create_date as dialDate, count(id) as calls, sum(duration) as duration,
                sum(client_cost) as clientCost, sum(vendor_cost) as vendorCost,
                count(case when action = 'DNC' then 1 end) as dontCall,
                count(case when action = 'AM' then 1 end) as am,
                count(case when action in ('SIP',
                                'Hangup','Phone') then 1 end) as transfer,
                count(case when disposition = 'ANSWER' then 1 end
                     ) as answer,
                count(case when disposition = 'CONGESION' then 1 end) as congesion,
                count(case when disposition = 'BUSY' then 1 end) as busy,
                count(case when disposition = 'CANCEL' then 1 end) as cancel,
                count(case when disposition = 'NO ANSWER' then 1 end) as noAnswer,
                count(case when disposition is NULL then 1 end) as other,
                count(case when disposition = 'FAILED' then 1 end) as failed
            from  ",reportTable 
                         );
SET @SELECTWHERE = CONCAT('
            where  `campaign_id` =', campaignId 
                         );
SET @SELECTQUERY = CONCAT(@SELECTFROM,
                @SELECTWHERE 
                         );
SET @QUERY = CONCAT(@SELECTQUERY, '
            group by  `dialDate`
            order by  `dialDate` desc
            limit  3'
                   );
PREPARE stmt
    FROM  @QUERY;
EXECUTE stmt;
DEALLOCATE PREPARE stmt;
END

如果有任何解决方法，请给我建议......

我也试过索引，但它没有给我正确的结果...

Answer 1

首先找到 campaignId 的第三个最新 dialDate。然后在WHERE子句中加入AND dialDate >= @third

还有这个复合索引：INDEX(campaignId, dialDate)

如果这还不够，请提供

生成的查询。（我可能很难跟上连接。）
EXPLAIN SELECT ...

使用过程根据列中的状态计算大数据

Count large data based on status in the columns using procedure

stored-procedures

count

large-data

mariadb