Let's assume you have a table named T
, with a column named MyString
, which stores your JSON values and is typed as string
(such a table is defined below for the example).
- You'll first need to invoke
parse_json()
on your column (unless it's already typed as dynamic
and not as string
, in which case you can skip this step).
- Then you can access the
Date
property in your JSON value and use todatetime()
to cast it to type datetime
.
- Afterwards, you can filter by your external parameter (
dateTimeLowerBound
in the example below).
- Lastly - you can
project
the relevant properties you're interested in (Message
is in the first element in the Errors
array, and SourceSystemId
), and you can cast them to their expected types (e.g. long
and string
using tolong()
and tostring()
respectively).
Here's the example:
let dateTimeLowerBound = datetime(2017-01-21);
let T = datatable(MyString:string) // this table is just for the example
[
'{"Status": 2, "SourceSystemId": "4", "RequisitionId": null, "Errors": [ { "Code": "8002", "Message": "some message", "FieldName": "VendorNumber", "PartNumber": null }, { "Code": "8003", "Message": "", "FieldName": "PartNumber", "PartNumber": "" } ], "SuppName": "SomeSupp", "Date":"2017-02-22"}'
];
T
| project MyJson = parse_json(MyString)
| where todatetime(MyJson.Date) > dateTimeLowerBound
| project SourceSystemId = tolong(MyJson.SourceSystemId), Message = tostring(MyJson.Errors[0].Message)
The output of this example should be a table with 2 columns, named SourceSystemId
and Message
, of types long
and string
, and with the values 4
an some message
, respectively.