sql - 在SQL Server中如何从不在另一个表中的每个组中获取最大时间戳

Question

在 SQL Server 中，我有下面的表结构

alert_details_status表:

alertTypeId macId     timeStamp     is_notified  escalation_status  alertType
123         13446   1547722123000       true       completed          408

alert_details 表:

basicid alertType   alertTypeId alertDescription    macId     timeStamp   companyId     alertName            alertCondition  unitType   channelType  alertValue expectedValue
1234     406            123         testalert       13446   1547722123000   1234     test alert name            testalert   Centimeters   length        50          60
1295     409            127         testalert       13448   1547722123000   1234     test alert name            testalert   Centimeters   length        50.2        60.3
1298     409            128         testalert       13448   1547722123000   1234     test alert name            testalert   Centimeters   length        50.2        60.3
1238     408            123         testalert       13446   1548148705000   1234     test alert name            testalert   Centimeters   length        50.2        60.3
1255     409            128         testalert       13448   1548135899000   1234     test alert name            testalert   Centimeters   length        50.2        60.3
1256     409            128         testalert       13448   1548135899000   1234     test alert name            testalert   Centimeters   length        50.2        60.3

我正在尝试对具有最大时间戳的alertType、alertTypeId、macid进行分组(应该只返回一个每组的数据(如果时间戳也相同)来自alert_details，但不在alert_details_status中。在两个表中，alertType、alertTypeId、macid 应该不同，然后每个组只应显示具有最大时间戳的alert_details 的结果。

我正在使用以下查询

SELECT * 
FROM 
    (SELECT 
         *, 
         ROW_NUMBER() OVER (PARTITINO BY alerttype, alerttypeid, macid ORDER BY timestamp DESC) RN 
     FROM 
         [test].[dbo].[alertdetails]) a 
WHERE 
    rn = 1 
    AND NOT EXISTS (SELECT macId, alertTypeId, MAX(timestamp) 
                    FROM [test].[dbo].[alert_details_status] AS c  
                    WHERE a.macId = c.macId 
                      AND a.alertTypeId = c.alertTypeId 
                      AND a.alertType = c.alertType 
                      AND (is_notified = 'true' AND escalation_status = 'completed') 
                   GROUP BY alertTypeId, macId)

但它返回错误的数据。

我想要的最终数据是

basicid alertType   alertTypeId alertDescription    macId     timeStamp   companyId     alertName            alertCondition  unitType   channelType  alertValue expectedValue
1295     409            127         testalert       13448   1547722123000   1234     test alert name            testalert   Centimeters   length        50.2        60.3
1238     408            123         testalert       13446   1548148705000   1234     test alert name            testalert   Centimeters   length        50.2        60.3
1256     409            128         testalert       13448   1548135899000   1234     test alert name            testalert   Centimeters   length        50.2        60.3

Answer 1

您首先计算然后过滤数据，这可能是问题所在。您可以尝试先过滤数据，然后再计算。

您可以尝试如下查询。

SELECT * 
FROM   (SELECT t1.*, 
               Row_number() 
                 OVER( 
                   partition BY t1.alerttype, t1.alerttypeid, t1.macid 
                   ORDER BY t1.timestamp DESC) rn 
        FROM   alert_details 
               LEFT JOIN alert_details_status t2 
                      ON macid = c.macid 
                         AND t1.alerttypeid = t2.alerttypeid 
                         AND t1.alerttype = t2.alerttype 
                         AND t2.is_notified = true 
                         AND t2.escalation_status = 'completed' 
        WHERE  t2.alerttypeid IS NULL) t 
WHERE  t.rn = 1