Using top-nested but not sure if I'm doing it right

[Tom Morgan]

Hi,

Hope this is an OK place for a Kusto Query language question...

Given some data, where there might be multiple versions of the same data, for instance:

+----+---------+-------+-------+
| ID | Version | Attr1 | Attr2 |
+----+---------+-------+-------+
|  1 |       1 |   123 |  10.0 |
|  2 |       1 |   123 |   5.0 |
|  2 |       2 |   222 |   4.0 |
|  3 |       1 |   123 |   5.0 |
|  3 |       2 |   222 |   4.0 |
|  3 |       3 |   333 |   5.0 |
+----+---------+-------+-------+

And, what I want to return are only the latest versions of all the rows, eg:

+----+---------+-------+-------+
| ID | Version | Attr1 | Attr2 |
+----+---------+-------+-------+
|  1 |       1 |   123 |  10.0 |
|  2 |       2 |   222 |   4.0 |
|  3 |       3 |   333 |   5.0 |
+----+---------+-------+-------+

Right, now, to achieve that I'm doing:

TableName
| top-nested of ID by max(1) , top-nested 1 of Version by max(Version), top-nested of Attr1 by max(1), top-nested of Attr2 by max(1)

It works, but it's not very pretty, and in the real-world there are a lot more columns, so it gets big quick.

Is there a better way to achieve what I'm trying to do?

[#2 · Accepted solution]

Try using "summarize arg_max()" : https://docs.microsoft.com/en-us/azure/kusto/query/arg-max-aggfunction

[#3]

@Yoni  - thank you!  Works perfectly.  For everyone else, this achieves the same results as my original one, but much cleaner and faster:

TableName
| summarize arg_max(Version,*) by ID