UpdateRecord makes use of the NiFi RecordPath Domain-Specific Language (DSL) to allow the user to indicate which field(s) in the Record should be updated. Users do this by adding a User-defined Property to the Processor's configuration. The name of the User-defined Property must be the RecordPath text that should be evaluated against each Record. The value of the Property specifies what value should go into that selected Record field.

When specifying the replacement value (the value of the User-defined Property), the user is able to specify a literal value such as the number 10; an Expression Language Expression to reference FlowFile attributes, such as ${filename}; or another RecordPath path from which to retrieve the desired value from the Record itself. Whether the value entered should be interpreted as a literal or a RecordPath path is determined by the value of the <Replacement Value Strategy> Property.

If a RecordPath is given and does not match any field in an input Record, that Property will be skipped and all other Properties will still be evaluated. If the RecordPath matches exactly one field, that field will be updated with the corresponding value. If multiple fields match the RecordPath, then all fields that match will be updated. If the replacement value is itself a RecordPath that does not match, then a null value will be set for the field. For instances where this is not the desired behavior, RecordPath predicates can be used to filter the fields that match so that no fields will be selected. See RecordPath Predicates for more information.

Below, we lay out some examples in order to provide clarity about the Processor's behavior. For all of the examples below, consider the example to operate on the following set of 2 (JSON) records:

  [{
    "id": 17,
    "name": "John",
    "child": {
        "id": "1"
    },
    "siblingIds": [4, 8],
    "siblings": [
        { "name": "Jeremy", "id": 4 },
        { "name": "Julia", "id": 8 }
    ]
  },
  {
    "id": 98,
    "name": "Jane",
    "child": {
        "id": 2
    },
    "gender": "F",
    "siblingIds": [],
    "siblings": []
  }]

For brevity, we will omit the corresponding schema and configuration of the RecordReader and RecordWriter. Otherwise, consider the following set of Properties are configured for the Processor and their associated outputs.

Example 1 - Replace with Literal

Here, we will replace the name of each Record with the name 'Jeremy' and set the gender to 'M':

This will yield the following output:

  [{
    "id": 17,
    "name": "Jeremy",
    "child": {
        "id": "1"
    },
    "gender": "M",
    "siblingIds": [4, 8],
    "siblings": [
        { "name": "Jeremy", "id": 4 },
        { "name": "Julia", "id": 8 }
    ]
  },
  {
    "id": 98,
    "name": "Jeremy",
    "child": {
        "id": 2
    },
    "gender": "M",
    "siblingIds": [],
    "siblings": []
  }]

Note that even though the first record did not have a "gender" field in the input, one will be added after the "child" field, as that's where the field is located in the schema.

Example 2 - Replace with RecordPath

This example will replace the value in one field of the Record with the value from another field. For this example, consider the following set of Properties:

This will yield the following output:

  [{
    "id": 17,
    "name": "Jeremy",
    "child": {
        "id": "1"
    },
    "siblingIds": [4, 8],
    "siblings": [
        { "name": "Jeremy", "id": 4 },
        { "name": "Julia", "id": 8 }
    ]
  },
  {
    "id": 98,
    "name": null,
    "child": {
        "id": 2
    },
    "gender": "F",
    "siblingIds": [],
    "siblings": []
  }]

Example 3 - Replace with Relative RecordPath

In the above example, we replaced the value of field based on another RecordPath. That RecordPath was an "absolute RecordPath," meaning that it starts with a "slash" character (/) and therefore it specifies the path from the "root" or "outer most" element. However, sometimes we want to reference a field in such a way that we defined the RecordPath relative to the field being updated. This example does just that. For each of the siblings given in the "siblings" array, we will replace the sibling's name with their id's. To do so, we will configure the processor with the following properties:

Note that the RecordPath that was given for the value starts with .., which is a reference to the parent. We do this because the field that we are going to update is the "name" field of the sibling. To get to the associated "id" field, we need to go to the "name" field's parent and then to its "id" child field. The above example results in the following output:

  [{
    "id": 17,
    "name": "John",
    "child": {
        "id": "1"
    },
    "siblingIds": [4, 8],
    "siblings": [
        { "name": "4", "id": 4 },
        { "name": "8", "id": 8 }
    ]
  },
  {
    "id": 98,
    "name": "Jane",
    "child": {
        "id": 2
    },
    "gender": "F",
    "siblingIds": [],
    "siblings": []
  }]

Example 4 - Replace Multiple Values

This example will replace the value of all fields that have the name "id", regardless of where in the Record hierarchy the field is found. The value that it uses references the Expression Language, so for this example, let's assume that the incoming FlowFile has an attribute named "replacement.id" that has a value of "91":

This will yield the following output:

  [{
    "id": 91,
    "name": "John",
    "child": {
        "id": "91"
    },
    "siblingIds": [4, 8],
    "siblings": [
        { "name": "Jeremy", "id": 91 },
        { "name": "Julia", "id": 91 }
    ]
  },
  {
    "id": 91,
    "name": "Jane",
    "child": {
        "id": 91
    },
    "gender": "F",
    "siblingIds": [],
    "siblings": []
  }]

It is also worth noting that in this example, some of the "id" fields were of type STRING, while others were of type INT. This is okay because the RecordReaders and RecordWriters should handle these simple type coercions for us.

Example 5 - Use Expression Language to Modify Value

This example will capitalize the value of all 'name' fields, regardless of where in the Record hierarchy the field is found. This is done by referencing the 'field.value' variable in the Expression Language. We can also access the field.name variable and the field.type variable.

This will yield the following output:

  [{
    "id": 17,
    "name": "JOHN",
    "child": {
        "id": "1"
    },
    "siblingIds": [4, 8],
    "siblings": [
        { "name": "JEREMY", "id": 4 },
        { "name": "JULIA", "id": 8 }
    ]
  },
  {
    "id": 98,
    "name": "JANE",
    "child": {
        "id": 2
    },
    "gender": "F",
    "siblingIds": [],
    "siblings": []
  }]