Testing

Now let's remove our old snapshots and re-run the test suite.

Copy
rm -rf snapshots

And re-run the tests with the snapshots directory:

Copy
rm -rf snapshots

Copy
ndc-test test --endpoint http://0.0.0.0:8100 --snapshots-dir snapshots

OR

Copy
cargo run --bin ndc-test -- test --endpoint http://localhost:8100 --snapshots-dir snapshots

We can see that predicate tests are passing, but some other test cases are not. That's okay - we'll keep iterating over the next few videos until we have all green tests here.

In our snapshots directory, we can also see that we're returning the correct data for some simple predicate queries. This query searches for albums for artist with ID 5, and we can see that the response contains the correct rows.

Request:

Copy
{
  "collection": "albums",
  "query": {
    "fields": {
      "artist_id": {
        "type": "column",
        "column": "artist_id"
      },
      "id": {
        "type": "column",
        "column": "id"
      },
      "title": {
        "type": "column",
        "column": "title"
      }
    },
    "limit": 10,
    "where": {
      "type": "binary_comparison_operator",
      "column": {
        "type": "column",
        "name": "artist_id",
        "path": []
      },
      "operator": {
        "type": "equal"
      },
      "value": {
        "type": "scalar",
        "value": 5
      }
    }
  },
  "arguments": {},
  "collection_relationships": {}
}

Response:

Copy
[
  {
    "rows": [
      {
        "artist_id": 1,
        "id": 1,
        "title": "For Those About To Rock We Salute You"
      },
      {
        "artist_id": 2,
        "id": 2,
        "title": "Balls to the Wall"
      },
      {
        "artist_id": 2,
        "id": 3,
        "title": "Restless and Wild"
      },
      {
        "artist_id": 1,
        "id": 4,
        "title": "Let There Be Rock"
      },
      {
        "artist_id": 3,
        "id": 5,
        "title": "Big Ones"
      },
      {
        "artist_id": 4,
        "id": 6,
        "title": "Jagged Little Pill"
      },
      {
        "artist_id": 5,
        "id": 7,
        "title": "Facelift"
      },
      {
        "artist_id": 6,
        "id": 8,
        "title": "Warner 25 Anos"
      },
      {
        "artist_id": 7,
        "id": 9,
        "title": "Plays Metallica By Four Cellos"
      },
      {
        "artist_id": 8,
        "id": 10,
        "title": "Audioslave"
      }
    ]
  }
]

Now let's deploy to Hasura and see how the GraphQL schema looks. Check out the video to follow along here.

Let's add a where clause to fetch the albums for artist ID 1.

We can see that the generated SQL is correct and that we're correctly parameterizing the query as well.

Let's try a query which uses a logical operator in the Hasura DDN GraphiQL console:

Copy
query MyQuery {
  albums(where: {
    _or: [
      {artist_id: {_eq: 1}}, 
      {artist_id: {_eq: 2}}, 
      {artist_id: {_eq: 5}}
    ]
  }) {
    title
  }
}

Again, we generate valid SQL and parameters, although we do have too many parentheses here. That's something we can improve later, but it's better to err on the safe side for now.

We've added support for basic where clauses, and we'll come back and fill in some of the missing expression types later, but next time, we'll take a look at order by clauses.