-1

我有一个数据库，其中包含一组带有用户 ID 和时间戳的事件，我正在尝试编写一个查询，该查询将为我提供每天触发事件的不同用户的计数。因此，如果我们有以下数据：

Event   | UID | Time Stamp
event 1 |  0  | 9/25/19 9:00 AM 
event 2 |  1  | 9/25/19 3:00 PM 
event 3 |  2  | 9/26/19 2:00 PM 
event 4 |  1  | 9/28/19 5:00 PM 
event 5 |  3  | 9/29/19 7:00 AM

那么输出应该是：

9/25/19 : 2
9/26/19 : 3
9/27/19 : 3 (since there are no new events on the 27th)
9/28/19 : 3 (since user with UID=1 has already been counted)
9/29/19 : 4

我有一个查询，它将获取每天的事件数，而不是直到那一天的所有日子的事件数。任何帮助将不胜感激！

4

1 回答 1

1

Kusto/ADX 中有几个内置的用户分析插件：https ://docs.microsoft.com/en-us/azure/kusto/query/useranalytics

例如，其中之一是activity_engagement插件：https ://docs.microsoft.com/en-us/azure/kusto/query/activity-engagement-plugin

例如：

let T = datatable(Event:string, UID:int, Timestamp:datetime)
[
    'event 1', 0,  datetime(9/25/19 9:00 AM),
    'event 2', 1,  datetime(9/25/19 3:00 PM), 
    'event 3', 2,  datetime(9/26/19 2:00 PM),
    'event 4', 1,  datetime(9/28/19 5:00 PM),
    'event 5', 3,  datetime(9/29/19 7:00 AM),
]
;
let min_date_time = toscalar(T | summarize startofday(min(Timestamp)));
let max_date_time = toscalar(T | summarize startofday(max(Timestamp)));
T
| evaluate activity_engagement (UID, Timestamp, 1d, 1d + max_date_time - min_date_time)
| project Timestamp, dcount_activities_outer

而且，如果您想在 9 月 27 日“填补空白”，您可以执行以下操作：

let T = datatable(Event:string, UID:int, Timestamp:datetime)
[
    'event 1', 0,  datetime(9/25/19 9:00 AM),
    'event 2', 1,  datetime(9/25/19 3:00 PM), 
    'event 3', 2,  datetime(9/26/19 2:00 PM),
    'event 4', 1,  datetime(9/28/19 5:00 PM),
    'event 5', 3,  datetime(9/29/19 7:00 AM),
]
;
let min_date_time = toscalar(T | summarize startofday(min(Timestamp)));
let max_date_time = toscalar(T | summarize startofday(max(Timestamp)));
range Timestamp from min_date_time to max_date_time step 1d
| join kind=leftouter (
    T
    | evaluate activity_engagement (UID, Timestamp, 1d, 1d + max_date_time - min_date_time)
    | project Timestamp, dcount_activities_outer
) on Timestamp
| order by Timestamp asc
| project Timestamp, coalesce(dcount_activities_outer, prev(dcount_activities_outer))
于 2019-09-30T20:59:30.087 回答