Set Reader and Writer Configuration Properties

DataWeave provides configuration properties for data formats, such as JSON (application/json), XML (application/xml), and (application/csv). The properties change the behavior of DataWeave readers and writers for those formats. For example, the default separator for a CSV reader is a comma (,). You can use the format’s separator property to specify a different separator for CSV content.

Before you begin, note that DataWeave version 2 (%dw 2.0) is for Mule 4 apps. For a Mule 3 app, refer to DataWeave version 1 (%dw 1.0) examples, within the Mule 3.9 documentation set. For other Mule versions, you can use the Mule Runtime version selector in the table of contents.

Use a Writer Property in an Output Directive

The following example shows how to append writer properties to the DataWeave output directive. The script uses indent = false to compress the JSON output into a single line.

DataWeave Script:

%dw 2.0
output application/json indent = false
---
{
	hello : "world",
	bello : "world",
	mello : "world"
}

Output JSON:

{"hello": "world","bello": "world","mello": "world"}

The following examples also append writer configuration properties to the output directive:

Avro example that uses schemaUrl
Excel (XLSX) example that uses header=true
Flat File example that sets a schemaPath
XML example that uses inlineCloseOn="empty" to close any empty XML elements

Use Reader and Writer Properties in DataWeave Functions

The DataWeave read, readUrl, and write functions accept one or more comma-separated property configurations within curly braces.

In the header of the following script, the value of myVar is a read function that inputs an XML sample with an empty child element (<ex1></ex1>). The function passes an XML reader property {nullValueOn: "empty"} that converts the value of the empty element to null.

In the body of the script, a write function accepts the value of myVar as input. The function passes the JSON writer properties {skipNullOn:"objects", writeAttributes:true} to skip the object with the null value (<ex1>null</ex1>) and to write the attribute and value of <ex3 a='greeting'>hello</ex3>.

DataWeave Script:

%dw 2.0
var myVar = read("<greeting><ex1></ex1><ex2>hello</ex2><ex3 a='greeting'>hello</ex3></greeting>", "application/xml", {nullValueOn: "empty"})
output application/json with binary
---
write(myVar.greeting, "application/json", {skipNullOn:"objects", writeAttributes:true})

Output JSON:

{
  "ex2": "hello",
  "ex3": {
    "@a": "greeting",
    "__text": "hello"
  }
}

The following examples pass reader properties to readUrl:

Use a Reader Property through a Connector Operation

Connector operations that read input data provide an outpuMimeType field that you can use to identify the input MIME type. This field also accepts one or more reader properties, which is useful for helping DataWeave read the input correctly.

By default, the DataWeave reader treats a comma (,) as a separator. However, assume that a CSV input is separated by the pipe (|) but that the content also contains commas. You can set a CSV reader property to recognize the pipe as the separator.

Pipe-separated CSV Input

id | name | role
1234 | Shokida,Mr. | dev
2345 | Achaval,Mr. | arch
3456 | Felisatti,Ms. | mngr
4567 | Chibana,Mr. | dev

The following example uses the Read operation in the File connector to read the pipe-separated (|) CSV input, and it uses a DataWeave script in the Transform Message component to output a row of the input in comma-separated format.

Mule Flow:

<flow name="ex-use-csv-reader-props" >
	<scheduler doc:name="Scheduler" >
		<scheduling-strategy >
			<fixed-frequency frequency="90" timeUnit="SECONDS"/>
		</scheduling-strategy>
	</scheduler>
	<file:read doc:name="Read" config-ref="File_Config" path="staff.csv"
	           outputMimeType='application/csv; separator=|; header=true'/>
	<ee:transform doc:name="Transform Message" >
		<ee:message>
			<ee:set-payload><![CDATA[%dw 2.0
output application/csv header = false
---
payload[1]]]></ee:set-payload>
		</ee:message>
	</ee:transform>
	<logger level="INFO" doc:name="Logger" message="#[payload]"/>
</flow>

The Scheduler component (<scheduler/>) in the example generates a Mule event each time it triggers the flow.

The Read operation (<file:read/>) uses outputMimeType='application/csv; separator=|; header=true' to identify the input MIME type (application/csv), the CSV separator (|), and the header setting (true).

The script in the Transform Message component (<ee:transform-message/>) returns the second row from the input CSV (2345 , Achaval\,Mr. , arch). The script uses payload[1] to select that row because DataWeave treats each row in the CSV as an index of an array. The output is comma-separated because the default CSV separator is a comma. The script omits the CSV header from the output by using the output directive (output application/csv header = false).

The Logger component (<logger/>), which is set to payload, writes the comma-separated output as INFO in its LoggerMessageProcessor message: 2345 , Achaval\,Mr. , arch. To avoid treating the comma before "Mr." as a separator, the CSV output escapes the comma with a backslash (Achaval\,Mr.).

The following examples pass reader properties through the outputMimeType field:

CSV example that uses streaming=true; header=true
JSON example that uses streaming=true

Use a Mule Variable as a Configuration Value

The following Mule flow transforms comma-separated CSV data into pipe-separated (|) CSV data. Instead of setting a literal value in the script, the example selects the pipe value from a Mule variable, which is part of a Mule event that travels through data-processing components in the flow.

The Scheduler component (<scheduler/>) in the example generates a Mule event each time it triggers the flow.

The Set Variable component (<set-variable/>) creates a Mule variable named delimiter with the value '|'. In Studio, the value is created as an fx expression (value="#['|']"), rather than a simple string (value="'|'").

The Transform Message component (<ee:transform-message/>) contains a DataWeave script that reads comma-separated CSV input and uses the Mule variable to write pipe-separated CSV output.

The Logger component (<logger/>), which is set to payload, writes the pipe-separated output as INFO in its LoggerMessageProcessor message: macaroni | rigatoni | ravioli | spaghetti.

Mule Flow:

<flow name="ex-use-mule-var-as-prop-config-val" >
  <scheduler doc:name="Scheduler" >
    <scheduling-strategy >
      <fixed-frequency frequency="45" timeUnit="SECONDS"/>
    </scheduling-strategy>
  </scheduler>
  <set-variable value="#['|']" doc:name="Set Variable" variableName="delimiter" />
  <ee:transform doc:name="Transform Message" >
    <ee:message>
      <ee:set-payload><![CDATA[%dw 2.0
var myVar = read("macaroni , rigatoni , ravioli , spaghetti", "application/csv" , {"header":false, separator:','})
output application/csv with binary
---
write(myVar, "application/csv", {"header":false, "separator":vars.delimiter})]]></ee:set-payload>
    </ee:message>
  </ee:transform>
  <logger level="INFO" doc:name="Logger" message="#[payload]"/>
</flow>

To provide comma-separated input for the example, the script in the Transform Message component uses a read function as a value to a DataWeave variable, myVar:

var myVar = read("macaroni , rigatoni , ravioli , spaghetti", "application/csv" , {"header":false, separator:','})

To write pipe-separated output, the DataWeave write function in the body of the script changes the comma-separated input to pipe-separated output. To select the pipe value from the Mule variable (delimiter), the function passes the writer property configuration "separator":vars.delimiter as an argument:

write(myVar, "application/csv", {"header":false, "separator":vars.delimiter})

The script uses the following DataWeave output directive in the script’s header: `output applicatio:page-aliases:

Note that if you want to see pipe-separated content from the Tranform Message component’s Preview screen in the Studio UI, you must create sample data for it:

Double-click the Transform Message component from the Message Flow area of the Studio app.
Right-click delimiter: String in the Context tab, which is located on the bottom left of the configuration UI for the component.
Select Edit Sample Data to open a vars-delimiter tab.
Replace any content in the vars-delimiter with the following sample data for the delimiter value: "|". You must put the delimiter in quotation marks.

For further guidance with sample data, see Previewing the Output of a Transformation.