如何在MySQL中獲取分組計數變量的最大值

Question

我有一個這樣的MySQL表；

recordID| netcall    | sign | activity | netid
1       | group1     | wa1  | 1        | 20
2       | group2     | wa2  | 2        | 30
3       | group1     | wa2  | 1        | 20
4       | group2     | wa3  | 2        | 30
5       | group1     | wa1  | 1        | 40
6       | group3     | wa4  | 3        | 50
7       | group3     | wa4  | 3        | 50
8       | group1     | wa2  | 1        | 40
9       | group1     | wa1  | 1        | 40
10      | group2     | wa4  | 2        | 60

我需要的是：

Netcall | count | activity | netid
Group1  | 3     | 1        | 40
Group2  | 2     | 2        | 30
Group3  | 2     | 3        | 50

我以為我可以；

SELECT MAX(xx.mycount) AS MAXcount
  FROM (SELECT COUNT(tt.sign) AS mycount ,tt.activity
        FROM NetLog tt
       WHERE ID <> 0
       GROUP BY netcall) xx

但這只會帶來未通過網絡電話細分的總數。 我沒有看到此問題的示例，但我確定有一個問題，我只是在問錯。

Answer 1

您的示例和期望輸出太基本了，您應該嘗試擴展，以便包括更多案例。

現在，您可以使用以下命令獲得期望的輸出：

SELECT `netcall`, COUNT(*) as `total`, MAX(`activity`) as `activity`
FROM t
GROUP BY `netcall`;

我的猜測是您可以在小組中進行不同的活動，因此您需要多個步驟

• 計算GROUP BY netcall, activity通話的COUNT() GROUP BY netcall, activity我稱之為q
• 然后查看每個我稱之為p通話的MAX(total)是netcall
• 現在，您將q作為o重用了q ，因此您只需擁有最大計數即可。

SQL演示

SELECT o.`netcall`, o.total, o.`activity`
FROM (
      SELECT `netcall`, COUNT(*) `total`, `activity`
      FROM t
      GROUP BY `netcall`, `activity`
     ) o 
JOIN (     
      SELECT `netcall`, MAX(`total`) as `total`
      FROM (
        SELECT `netcall`, COUNT(*) `total`
        FROM t
        GROUP BY `netcall`, `activity`
      ) q  
      GROUP BY `netcall`
     ) p
  ON o.`netcall` = p.`netcall`
 AND o.`total` = p.`total`

使用MySQL v8 +，您可以使用cte和window函數來簡化一點

with group_count as (
  SELECT `netcall`, COUNT(*) as total, `activity`        
  FROM t
  GROUP BY `netcall`, `activity`
), group_sort as (
  SELECT `netcall`, total, `activity`, 
          RANK() OVER (PARTITION BY `netcall`, `activity` ORDER BY total DESC) as rnk
  FROM group_count
)  
SELECT *
FROM group_sort
WHERE rnk = 1

Answer 2

每天都在SO上詢問（並回答）這個問題； 它甚至在MySQL手冊中都有自己的章節，但是無論如何...

SELECT a.netcall
     , b.total
     , a.activity
  FROM netlog a
  JOIN
     ( SELECT netcall 
            , MAX(record_id) record_id
            , COUNT(*) total
         FROM netlog 
        GROUP 
           BY netcall 
    ) b
   ON b.netcall = a.netcall
  AND b.record_id = a.record_id

Answer 3

    SELECT k.netcall, k.netID,  MAX(k.logins) highest, 
AVG(k.logins) average, netDate, activity
                          FROM
                            (SELECT netID, netcall, COUNT(*) logins, DATE(`logdate`) as netDate, activity
                               FROM NetLog
                              WHERE netID <> 0 AND status = 1
                                AND netcall <> '0' AND netcall <> ''
                              GROUP BY netcall, netID) k
                            GROUP BY netcall  
                            ORDER BY highest DESC 

結果：

Net Call    Highest Average Net ID  Sub Net Of...   ICS
214   309   Map Date    Activity
MESN    65  41.5294 339     214     309 MAP 2017-09-03  MESN
W0KCN   34  14.9597 1       214     309 MAP 2016-03-15  KCNARES Weekly 2m Voice Net
W0ERH   31  31.0000 883     214     309 MAP 2018-10-12  Johnson Co. Radio Amateurs Club Meeting Net
KCABARC 29  22.3333 57      214     309 MAP 2016-10-10  KCA Blind Amateurs Weekly 2m Voice Net

....