如何跨行对整数数组求和

Question

如何实现 summarize sum(Trend) by id 的等价物，其中 Trend 是整数数组？

Input:
——————————————————————————
Id    | ParentId | Trend 
——————————————————————————
C1-P1 | P1       | [1,2,3]
C2-P1 | P1       | [4,5,6]
C3-P1 | P1       | [1,1,1]
P1    |          |
C1-P2 | P2       | [4,5,6]
C2-P2 | P2       | [7,8,9]
P2    |          |
—————————————————————————-

Needed Output:
——————————————————————————
Id    | ParentId | Trend 
——————————————————————————
C1-P1 | P1       | [1,2,3]
C2-P1 | P1       | [4,5,6]
C3-P1 | P1       | [1,1,1]
P1    |          | [6,8,10] 
C1-P2 | P2       | [4,5,6]
C2-P2 | P2       | [7,8,9]
P2    |          | [11,13,15]
—————————————————————————-

Answer 1

请检查以下查询是否解决了您的情况：它使用 mv-expand operator and 'with_itemindex' option 来扩展数组的值。

let _data = datatable(Id:string, ParentId:string, Trend:dynamic)
[
'C1-P1','P1', dynamic([1,2,3]),
'C2-P1', 'P1',  dynamic([4,5,6]),
'C3-P','P1',dynamic([1,1,1]),
'P1','',dynamic([]),
'C1-P2','P2',dynamic([4,5,6]),
'C2-P2','P2',dynamic([7,8,9]),
'P2', '', dynamic([])
];
_data 
| mv-expand with_itemindex=x Trend to typeof(long)
| summarize sum(Trend) by  ParentId, x
| summarize Trend=make_list(sum_Trend) by ParentId
| union (_data | where isnotempty( ParentId))

如何跨行对整数数组求和

How to sum integer arrays across rows

azure-data-explorer

kql