Json

The JsonSource and JsonDestination allows you to read or write data from/into a json format, either into a file or a webservice.

Json connector package#

The ETLBox.Json   package include the JsonSource and JsonDestination classes, along with a JsonPathConverter and an ExpandoJsonPathConverter. The json source can send json received from a file or from a web endpoint into a dataflow. The json destination can either create a json file or send the json into any kind of web service. This class is ideal if you need to establish connections with web services, e.g. an REST API endpoint or the result of a GraphQL query.

The json connector package is based on Json.NET by Newtonsoft   .

If you want to start with example code right away, you will find it in the recipes section for the JsonSource and JsonDestination. The components could also be used in other examples.

JsonSource#

Json Source let you read a json.

Here an example for a response from a json web service request when sending a GET to the url http://test.com

[
  {
    "Col1": 1,
    "Col2": "Test1"
  },
  {
    "Col1": 2,
    "Col2": "Test2"
  },
  {
    "Col1": 3,
    "Col2": "Test3"
  }
]

Here is some code that would read this json response, and deserialize the json into the object MySimpleRow.

public class MySimpleRow
{
    public int Col1 { get; set; }
    public string Col2 { get; set; }
}

JsonSource<MySimpleRow> source = new JsonSource<MySimpleRow>("http://test.com/", ResourceType.Http);
MemoryDestination<MySimpleRow> dest = new MemoryDestination<MySimpleRow>();

source.LinkTo(dest); //Link the components
Network.Execute(source); //Run the data flow

This code would then read the three entries from the source and post it into the MemoryDestination.

Nested arrays#

The array doesn’t need to be the top node of your json - it could be nested in your json file. Like this:

{
    "data": {
        "array": [
            {
                "Col1": 1,
                "Col2": "Test1"
            },
            ...
        ]
    }
}

ETLBox automatically scans the incoming json file and starts reading (and deserializing) after the first occurrence of the begin of an array (which is the “[” symbol).

Working with JsonSerializer#

Sometimes you have a more complex json structure. Here an example:

[
    {
        "Column1": 1,
        "Column2": {
            "Id": "A",
            "Value": "Test1"
        }
    },
    ...
]

In order to read map this json with an object, you would need to create two object which are associated with each other:

public class MyRow
{
    public int Column1 { get; set; }
    public MyIdValueObject Column2 { get; set; }
}

public class MyIdValueObject
{
    public string Id { get; set; }
    public string Value { get; set; }
}

JsonConverter#

Sometimes you don’t want to specify all objects that would map your json structure. To get around this the underlying JsonSerializer object that is used for deserialization is exposed by the JsonSource. See also the JsonSerializer docs from Newtonsoft.Json   for more details. Now you could add your own JsonConverter so that you could use JsonPath within your JsonProperty attributes. To make your programmer’s live even easier, you can use the JsonPathConverter class which is part of the ETLBox.Json package. Here is an example how to use it:

[JsonConverter(typeof(JsonPathConverter))]
public class MySimpleRow
{
    [JsonProperty("Column1")]
    public int Col1 { get; set; }
    [JsonProperty("Column2.Value")]
    public string Col2 { get; set; }
}

JsonSource<MySimpleRow> source = new JsonSource<MySimpleRow>("res/JsonSource/NestedData.json", ResourceType.File);

The JsonProperty attribute does accept a json path expression which is resolved to read the values from the json. This will also work for nested arrays and the * operator.

JsonConverter with ExpandoObject#

When you are working with dynamic object, it can be quite tricky to write a converter class for this. The ETLBox.Json comes with the class ExpandoJsonPathConverter which helps you to deserialize json into a dynamic objects using json path expressions.

Consider this your input data:

[
    {
        "Column1": 1,
        "Column2": {
            "Id": "A",
            "Value": "Test1"
        }
    }
]

You can parse this data by declaring your JsonSource and xpandoJsonPathConverter like this:

JsonSource<ExpandoObject> source = 
  new JsonSource<ExpandoObject>("http://test.com", ResourceType.Http);
List<JsonProperty2JsonPath> pathLookups = 
new List<JsonProperty2JsonPath>()
  {
      new JsonProperty2JsonPath()
      {
          SearchPropertyName = "Column2",
          JsonPath = "$.Value",
          OutputPropertyName = "Value"
        },
      new JsonProperty2JsonPath() {
          SearchPropertyName = "Column2",
          JsonPath = "$['Id']",
          OutputPropertyName = "Id"
        }
  };
source.JsonSerializer.Converters.Add(new ExpandoJsonPathConverter(pathLookups));

JsonDestination#

The result of your pipeline can be written as json using a JsonDestination.

The following code:

 public class MySimpleRow
{
    public string Col2 { get; set; }
    public int Col1 { get; set; }
}

JsonDestination<MySimpleRow> dest = 
  new JsonDestination<MySimpleRow>("test.json", ResourceType.File);

would result in the following json:

[
  {
    "Col1": 1,
    "Col2": "Test1"
  },
  {
    "Col1": 2,
    "Col2": "Test2"
  },
  {
    "Col1": 3,
    "Col2": "Test3"
  }
]

Like the JsonSource you can modify the exposed JsonSerializer to modify the serializing behavior. See the `JsonSerializer` docs in Newtonsoft.Json   for more details.

Using Json with arrays#

If you don’t want to use objects, you can use arrays with your JsonSource or JsonDestination. To use the source with a string array, your code would look like this:

JsonSource<string[]> source = new JsonSource<string[]>("https://test.com", ResourceType.Http);

Now you either have to override the JsonSerializer yourself in order to properly convert the json into a string[]. Or your incoming Json has to be in following format:

[
    [
        "1",
        "Test1"
    ],
    //etc.
]

When using the JsonDestination with a string array, the serialized data would look like the json above.

Working with dynamic objects#

JsonSource and destination support the usage of dynamic object. This allows you to use a dynamic ExpandoObject instead of a POCO. Here is an example for the source:

JsonSource<ExpandoObject> source = 
  new JsonSource<ExpandoObject>("source.json", ResourceType.File);

RowTransformation<ExpandoObject> trans = 
  new RowTransformation<ExpandoObject>(
    row =>
    {
        dynamic r = row as ExpandoObject;
        r.Col1 = r.Value1;
        r.Col2 = r.Value2;
        return r;
    });
DbDestination<ExpandoObject> dest = 
  new DbDestination<ExpandoObject>(Connection, "DynamicJson");

source.LinkTo(trans).LinkTo(dest);
Network.Execute(source);