我有这样的数据：
Date_time OS Login Action Download Action_detail Download_detail
09-09-19 Android User_id Download 50 Download - Games 20
09-09-19 Android User_id Download 50 Download - App 15
09-09-19 Android User_id Download 50 Download - Music 30
09-09-19 Android User_id Download 50 Download - Others 20
09-09-19 Android Guest Download 20 Download - Games 20
09-09-19 iOS User_id Download 10 Download - Others 15
09-09-19 iOS Guest Uninstall 15 Download - Games 20
09-09-19 iOS Guest Uninstall 15 Download - App 5
10-09-19 iOS Guest Uninstall 20 Download - App 15
10-09-19 iOS Guest Uninstall 20 Download - App 10
无论如何，我可以通过以下方式返回查询：
Date_time OS Login Action Download Action_detail Download_detail
09-09-19 Android User_id Download 50 Download - Games 20
09-09-19 Android User_id Download (null) Download - App 15
09-09-19 Android User_id Download (null) Download - Music 30
09-09-19 Android User_id Download (null) Download - Others 20
09-09-19 Android Guest Download 20 Download - Games 20
09-09-19 iOS User_id Download 10 Download - Others 15
09-09-19 iOS Guest Uninstall 15 Download - Games 20
09-09-19 iOS Guest Uninstall (null) Download - App 5
10-09-19 iOS Guest Uninstall 20 Download - App 15
10-09-19 iOS Guest Uninstall (null) Download - App 10
“下载编号”列是“下载操作”的唯一编号， 最后，我要用独特的date_time，OS，Login，Action作为切片器对其进行总结。
答案 0 :(得分：0)
您可以使用LAG功能获得所需的结果，如下所示-
SELECT Date_time,OS,Login,Action,Download,Action_detail,Download_detail,
CASE
WHEN Download = LAG(Download,1) OVER (ORDER BY (SELECT NULL FROM DUAL)) THEN NULL
ELSE Download
END New_Column
FROM your_table
WINDOW函数的ORDER存在问题，因为您可以看到我使用“ SELECT NULL FROM DUAL”来保持数据的常规ORDER。但这并不能确认以这种方式数据的顺序将始终相同。如果表中有任何列可用于订购，那将是正确的方法。