Flow variables
flowVars
-
Example:
#[flowVars.myVar]
In Mule 4, the DataWeave expression language replaces Mule Expression Language (MEL). This section provides guidance on migrating MEL expressions in your Mule apps to DataWeave (2.0 and later).
DataWeave is a Mule-specific language that can be used lightly as an expression language to simply reference values and evaluate conditions, or as a scripting language to construct complex data transformations that include functions and recursion.
Unlike MEL, it supports different data types out of the box, like JSON, XML, or CSV, which means data no longer needs to be transformed before expressions can be evaluated. It also supports advanced use cases like mapping functions or filtering data.
It’s important to notice that the main data selector in DataWeave is the .
operator.
Since it shares most keywords with MEL, the syntax for most scenarios is
identical. For example, #[message.payload.name]
is the same in both languages.
The main difference between MEL and DataWeave is that DataWeave expressions have no side effects. You can use DataWeave to extract or generate data but not to modify it. The Java Integration examples explain how to deal with those use cases.
The next sections show how to adapt some uses of MEL to Mule 4.
Except for the following changes, context variables (also called Mule Runtime variables) remain the same in DataWeave:
MEL | DataWeave |
---|---|
Flow variables
|
Mule Event Variables
|
Exception
|
|
Inbound Attachments
|
|
Inbound Properties
|
Mule Message Attributes (metadata)
|
Outbound Attachments
|
|
Outbound Properties
|
|
|
|
|
Removed |
Record Variables
|
Mule Event Variables
|
Session Variables
|
|
|
Removed: Use the DataWeave |
You can use DataWeave Selectors to query data as you did with MEL. Only now, you won’t need extra transformations when dealing with types like JSON.
The next examples log the ID from this JSON payload:
{
"name" : "Rachel Duncan",
"id": "779H41"
}
<json:json-to-object-transformer returnClass="java.util.Map" />
<logger message="Updating health check record for subject '#[payload.id]'" level="INFO" />
<logger message="Updating health check record for subject '#[payload.id]'" />
As mentioned earlier, DataWeave does not support modifying data in the way the expression-component
allowed with MEL. To achieve those use cases, you must use the Scripting module.
Consider payload described above as a Java Map. The examples modify the name
attribute.
<expression-component>
<![CDATA[
payload.name = 'R. Duncan'
]]>
</expression-component>
<script:execute engine="groovy">
<script:code>
payload.put('name', 'R. Duncan')
return payload
</script:code>
</script:execute>
To use the Scripting module, simply add it to your app using the Studio palette, or add the following dependency in your pom.xml
file:
<dependency>
<groupId>org.mule.modules</groupId>
<artifactId>mule-scripting-module</artifactId>
<version>1.1.0</version> <!-- or newer -->
<classifier>mule-plugin</classifier>
</dependency>
DataWeave can be used to invoke static methods. For regular methods, you can use the Java module, which exposes both an operation and a function to invoke Java methods.
The following examples use Java’s Locale getDefault
static method in Mule 3 and 4.
<validation:is-time time="#[payload]" pattern="h:mm a" locale="#[java.util.Locale.getDefault().getLanguage()]"/>
<validation:is-time time="#[payload]" pattern="h:mm a" locale="#[java!java::util::Locale::getDefault().getLanguage()]"/>
Note that DataWeave requires the java!
prefix to indicate that a Java static method is to be searched, and
it requires the fully qualified Class name (separated by ::
instead of .
).
The following examples show the usage of Java’s String equalsIgnoreCase
method in Mule 3 and 4.
<choice>
<when expression="#[payload.equalsIgnoreCase('error')]">
<logger message="An error message has been received." level="ERROR"/>
</when>
<otherwise>
<logger message="Message received: #[payload]" level="INFO"/>
</otherwise>
</choice>
<choice doc:name="Choice">
<when expression="#[Java::invoke('java.lang.String', 'equalsIgnoreCase(String)', payload, {arg0: 'error'})]">
<logger message="An error message has been received." level="ERROR"/>
</when>
<otherwise >
<logger message="Message received: #[payload]" level="INFO"/>
</otherwise>
</choice>
To use the Java module, simply add it to your app using the Studio palette, or add the following dependency to your pom.xml
file:
<dependency>
<groupId>org.mule.modules</groupId>
<artifactId>mule-java-module</artifactId>
<version>1.0.0</version> <!-- or newer -->
<classifier>mule-plugin</classifier>
</dependency>
The enricher
has been removed and replaced by target variables, which are now supported
by every operation. The example below shows how to send the result of an
operation to a variable in Mule 3 and 4.
<enricher target="#[flowVars.response]">
<http:request config-ref="HTTP_Request_Configuration" path="/get" method="GET"/>
</enricher>
<http:request config-ref="HTTP_Request_Configuration" path="/get" method="GET" target="response"/>
See:Enriching Data with Target Variables for more details.
DataWeave can be used to query XML content using its selectors, but you can also use the XML module, which exposes both an operation and a function to execute XPath queries.
The following examples show how to take lines containing a specific word taken from a variable in Mule 3 and 4.
<set‐variable variableName="word" value="handkerchief"/>
<expression‐transformer>
xpath3('//LINE[contains(.,$word)]',payload,'NODESET')
</expression‐transformer>
<set‐variable variableName="word" value="handkerchief"/>
<set-payload value="#[XmlModule::xpath('//LINE[contains(., \$word)]', payload, {'word': vars.word})]" />
To use the XML module, simply add it to your app using the Studio palette, or add the following dependency to your pom.xml
file:
<dependency>
<groupId>org.mule.modules</groupId>
<artifactId>mule-xml-module</artifactId>
<version>1.1.0</version> <!-- or newer -->
<classifier>mule-plugin</classifier>
</dependency>
DataWeave matching functions match
and matches
(see Core DataWeave Functions)
can be used instead. The next example shows how a regex is used in DataWeave to replace the use of the wildcard
MEL function.
<choice>
<when expression="#[wildcard('Hello *')]">
<set-payload value="Hello, how can I help?"/>
</when>
<otherwise>
<set-payload value="Courtesy requires a greeting."/>
</otherwise>
</choice>
<choice doc:name="Choice">
<when expression="#[payload matches /Hello\s[A-z]+/]">
<set-payload value="Hello, how can I help?"/>
</when>
<otherwise >
<set-payload value="Courtesy requires a greeting."/>
</otherwise>
</choice>