LEFT JOIN之前的MySQL计数字段值

Question

我一直试图让这个工作,但有些事情让我感到沮丧.我在这里搜索SO,但没有找到任何描述我具体案例的主题.我有以下架构

在这里小提琴:http://www.sqlfiddle.com/#!2/ac8162 / 6

表records:

| ID | CONTRACT |   NAME   | STATUS  |
|----|----------|----------|---------|
|  1 | US       | FOO      | status1 |
|  2 | US       | BAR      | status1 |
|  3 | UK       | ABC Inc. | status3 |
|  4 | PL       | EFG LTD. | status2 |
|  5 | UK       | XXX Inc. | status2 |
|----|----------|----------|---------|

表transactions:

| ID | RECORD_ID | RESPONSE_DELAY | STATUS  |
|----|-----------|----------------|---------|
|  1 |         1 |           8889 | status1 |
|  2 |         1 |           8813 | status1 |
|  3 |         1 |           5908 | status2 |
|  4 |         1 |           4779 | status3 |
|  5 |         2 |            519 | status1 |
|  6 |         2 |           8804 | status1 |
|  7 |         3 |           2604 | status1 |
|  8 |         3 |           5054 | status2 |
|  9 |         3 |            385 | status3 |
| 10 |         4 |           8852 | status1 |
| 11 |         4 |           4330 | status1 |
| 12 |         4 |           4507 | status1 |
| 13 |         4 |           6282 | status2 |
| 14 |         4 |           5789 | status3 |
| 15 |         4 |           8685 | status4 |
| 16 |         5 |           8339 | status1 |
| 17 |         5 |           9543 | status1 |
| 18 |         1 |           7870 | status1 |
|----|-----------|----------------|---------|

records有一对多的关系transactions.我试图通过合同分组报告.我使用以下查询:

SELECT 

t0.`contract` as contract,

COUNT(DISTINCT t0.`id`) as rec_num,
COUNT(t1.`id`) as tr_num,

AVG(t1.`response_delay`) as rd_avg,

SUM(CASE WHEN t0.`status`='status1' THEN 1 ELSE 0 END) as s1,
SUM(CASE WHEN t0.`status`='status2' THEN 1 ELSE 0 END) as s2,
SUM(CASE WHEN t0.`status`='status3' THEN 1 ELSE 0 END) as s3

FROM 
records as t0

LEFT JOIN transactions AS t1 
ON
(t0.id = t1.record_id)

GROUP BY contract ORDER BY t0.`id` DESC

结果:

| CONTRACT | REC_NUM | TR_NUM |  RD_AVG   | S1 | S2 | S3 |
|----------|---------|--------|-----------|----|----|----|
| PL       |       1 |      6 | 6407.5    |  0 |  6 |  0 |
| UK       |       2 |      5 | 5185      |  0 |  2 |  3 |
| US       |       2 |      7 | 6511.7143 |  7 |  0 |  0 |
|----------|---------|--------|-----------|----|----|----|

我的目标是计算每个合同的每种状态类型,只计算记录状态而不是事务状态,我追求的结果是:

| CONTRACT | REC_NUM | TR_NUM |  RD_AVG   | S1 | S2 | S3 |
|----------|---------|--------|-----------|----|----|----|
| PL       |       1 |      6 | 6407.5    |  0 |  1 |  0 |
| UK       |       2 |      5 | 5185      |  0 |  1 |  1 |
| US       |       2 |      7 | 6511.7143 |  2 |  0 |  0 |
|----------|---------|--------|-----------|----|----|----|

任何帮助是极大的赞赏

Answer 1

试试这个:

SELECT 

t0.`contract` as contract,

COUNT(DISTINCT t0.`id`) as rec_num,
COUNT(t1.`id`) as tr_num,

AVG(t1.`response_delay`) as rd_avg,

count(distinct CASE WHEN t0.`status`='status1' THEN t0.id  END) as s1,
count(distinct CASE WHEN t0.`status`='status2' THEN t0.id  END) as s2,
count(distinct CASE WHEN t0.`status`='status3' THEN t0.id  END) as s3


FROM 
records as t0

LEFT JOIN transactions AS t1 
ON
(t0.id = t1.record_id)


GROUP BY contract ORDER BY t0.`id` DESC