How to query Cosmos DB on Azure Portal with $v and $t

Question

I have a Cosmos DB on Azure with SQL API. I am able to query data through C# using their Nuget package, but I am getting an error when running a query from Azure Portal > Cosmos DB > Data Explorer.

I'd like to get the value of $v and do some filtering based on it.

Using

SELECT c.id, c.Remarks FROM c

I am getting results with nested objects (see the result below). But I need only one value from the nested object.

I tried changing the syntax to

SELECT c.id, c.Remarks.$v FROM c

or

SELECT c.id, c.Remarks.v FROM c

or

SELECT c.id, c.Remarks/$v FROM c

but I am getting an error.

Current Results:

[
    {
        "id": "e9f3ae8e47ab4bbca97dadf3ff1fe08c",
        "Remarks": {
            "$t": 2,
            "$v": "Success"
        }
    },
    {
        "id": "97bea2e9919c48f2bde83c11c50e8177",
        "Remarks": {
            "$t": 2,
            "$v": "Failure"
        }
    },
    {
        "id": "bb142e17b8184d5c84a21aa2e218e3be",
        "Remarks": {
            "$t": 2,
            "$v": "Success"
        }
    }
]

Expected Results:

I'd like to see (and preferably apply filter to only see only Failures)

[
    {
        "id": "e9f3ae8e47ab4bbca97dadf3ff1fe08c",
        "Remarks": "Success"
    },
    {
        "id": "97bea2e9919c48f2bde83c11c50e8177",
        "Remarks": "Failure"
    },
    {
        "id": "bb142e17b8184d5c84a21aa2e218e3be",
        "Remarks": "Success"
    }
]

I'm encountering errors like

Syntax error, invalid token '$'

but I cannot find any solution.

Answer 1

It looks like you're trying to use the SQL API against a document that was inserted via the Mongo endpoint. That's what causes the $t/$v stuff.

Mixing and matching Mongo and SQL APIs is a little complex and not really recommended.

You can do one of a few things, in no particular order.

1. Use Mongo to do queries
2. Recreate your collection using just SQL APIs
3. Checkout this answer about how to write SQL API queries against Mongo: Cosmos DB + pyDocumentDB + Databricks (You've gotta do little [] to access things with a $ at the beginning of them, like so: SELECT TOP 10 twttr["$v"].tweet["$v"].source FROM twttr )

I personally recommend not doing #3 unless you have to. If you want to use SQL, I'd recommend just using SQL and going with #2.