Talend Component Documentation

Talend Components Definitions Documentation

Components Definition

Talend Component framework relies on several primitive components.

They can all use @PostConstruct and @PreDestroy to initialize/release some underlying resource at the beginning/end of the processing.

in distributed environments class' constructor will be called on cluster manager node, methods annotated with @PostConstruct and @PreDestroy annotations will be called on worker nodes. Thus, partition plan computation and pipeline task will be performed on different nodes.

deployment diagram

  1. Created task consists of Jar file, containing class, which describes pipeline(flow) which should be processed in cluster.

  2. During partition plan computation step pipeline is analyzed and split into stages. Cluster Manager node instantiates mappers/processors gets estimated data size using mappers, splits created mappers according to the estimated data size. All instances are serialized and sent to Worker nodes afterwards.

  3. Serialized instances are received and deserialized, methods annotated with @PostConstruct annotation are called. After that, pipeline execution is started. Processor’s @BeforeGroup annotated method is called before processing first element in chunk. After processing number of records estimated as chunk size, Processor’s @AfterGroup annotated method called. Chunk size is calculated depending on environment the pipeline is processed by. After pipeline is processed, methods annotated with @PreDestroy annotation are called.

driver processing workflow

worker processing workflow

all framework managed methods MUST be public too. Private methods are ignored.
in term of design the framework tries to be as declarative as possible but also to stay extensible not using fixed interfaces or method signatures. This will allow to add incrementally new features of the underlying implementations.


A PartitionMapper is a component able to split itself to make the execution more efficient.

This concept is borrowed to big data world and useful only in this context (BEAM executions). Overall idea is to divide the work before executing it to try to reduce the overall execution time.

The process is the following:

  1. Estimate the size of the data you will work on. This part is often heuristic and not very precise.

  2. From that size the execution engine (runner for beam) will request the mapper to split itself in N mappers with a subset of the overall work.

  3. The leaf (final) mappers will be used as a Producer (actual reader) factory.

this kind of component MUST be Serializable to be distributable.


A partition mapper requires 3 methods marked with specific annotations:

  1. @Assessor for the evaluating method

  2. @Split for the dividing method

  3. @Emitter for the Producer factory


The assessor method will return the estimated size of the data related to the component (depending its configuration). It MUST return a Number and MUST not take any parameter.

Here is an example:

public long estimateDataSetByteSize() {
    return ....;

The split method will return a collection of partition mappers and can take optionally a @PartitionSize long value which is the requested size of the dataset per sub partition mapper.

Here is an example:

public List<MyMapper> split(@PartitionSize final long desiredSize) {
    return ....;

The emitter method MUST not have any parameter and MUST return a producer. It generally uses the partition mapper configuration to instantiate/configure the producer.

Here is an example:

public MyProducer create() {
    return ....;


Producer is the component interacting with a physical source. It produces input data for the processing flow.

A producer is a very simple component which MUST have a @Producer method without any parameter and returning any data:

public MyData produces() {
    return ...;


A Processor is a component responsible to convert an incoming data to another model.

A processor MUST have a method decorated with @ElementListener taking an incoming data and returning the processed data:

public MyNewData map(final MyData data) {
    return ...;
this kind of component MUST be Serializable since it is distributed.
if you don’t care much of the type of the parameter and need to access data on a "map like" based rule set, then you can use JsonObject as parameter type and Talend Component will just wrap the data to enable you to access it as a map. The parameter type is not enforced, i.e. if you know you will get a SuperCustomDto then you can use that as parameter type but for generic component reusable in any chain it is more than highly encouraged to use JsonObject until you have your an evaluation language based processor (which has its own way to access component). Here is an example:
public MyNewData map(final JsonObject incomingData) {
    String name = incomingData.getString("name");
    int name = incomingData.getInt("age");
    return ...;

// equivalent to (using POJO subclassing)

public class Person {
    private String age;
    private int age;

    // getters/setters

public MyNewData map(final Person person) {
    String name = person.getName();
    int name = person.getAge();
    return ...;

A processor also supports @BeforeGroup and @AfterGroup which MUST be methods without parameters and returning void (result would be ignored). This is used by the runtime to mark a chunk of the data in a way which is estimated good for the execution flow size.

this is estimated so you don’t have any guarantee on the size of a group. You can literally have groups of size 1.

The common usage is to batch records for performance reasons:

public void initBatch() {
    // ...

public void endBatch() {
    // ...
it is a good practise to support a maxBatchSize here and potentially commit before the end of the group in case of a computed size which is way too big for your backend.

Multiple outputs

In some case you may want to split the output of a processor in two. A common example is "main" and "reject" branches where part of the incoming data are put in a specific bucket to be processed later.

This can be done using @Output. This can be used as a replacement of the returned value:

public void map(final MyData data, @Output final OutputEmitter<MyNewData> output) {

Or you can pass it a string which will represent the new branch:

public void map(final MyData data,
                @Output final OutputEmitter<MyNewData> main,
                @Output("rejected") final OutputEmitter<MyNewDataWithError> rejected) {
    if (isRejected(data)) {
    } else {

// or simply

public MyNewData map(final MyData data,
                    @Output("rejected") final OutputEmitter<MyNewDataWithError> rejected) {
    if (isSuspicious(data)) {
        return createNewData(data); // in this case we continue the processing anyway but notified another channel
    return createNewData(data);

Multiple inputs

Having multiple inputs is closeto the output case excep it doesn’t require a wrapper OutputEmitter:

public MyNewData map(@Input final MyData data, @Input("input2") final MyData2 data2) {
    return createNewData(data1, data2);

@Input takes the input name as parameter, if not set it uses the main (default) input branch.

due to the work required to not use the default branch it is recommended to use it when possible and not name its branches depending on the component semantic.


An Output is a Processor returning no data.

Conceptually an output is a listener of data. It perfectly matches the concept of processor. Being the last of the execution chain or returning no data will make your processor an output:

public void store(final MyData data) {
    // ...


For now Talend Component doesn’t enable you to define a Combiner. It would be the symmetric part of the partition mapper and allow to aggregate results in a single one.

Configuring components

Component are configured through their constructor parameters. They can all be marked with @Option which will let you give a name to parameters (if not it will use the bytecode name which can require you to compile with -parameter flag to not have arg0, arg1, …​ as names).

The parameter types can be primitives or complex objects with fields decorated with @Option exactly like method parameters.

it is recommended to use simple models which can be serialized by components to avoid headaches when implementing serialized components.

Here is an example:

class FileFormat implements Serializable {
    private FileType type = FileType.CSV;

    private int maxRecords = 1024;

@PartitionMapper(family = "demo", name = "file-reader")
public MyFileReader(@Option("file-path") final File file,
                    @Option("file-format") final FileFormat format) {
    // ...

Using this kind of API makes the configuration extensible and component oriented letting the user define all he needs.

The instantiation of the parameters is done from the properties passed to the component (see next part).


What is considered as a primitive in this mecanism is a class which can be directly converted from a String to the expected type.

It obviously includes all java primitives, String type itself but also all the types with a org.apache.xbean.propertyeditor.Converter.

This includes out of the box:

  • BigDecimal

  • BigInteger

  • File

  • InetAddress

  • ObjectName

  • URI

  • URL

  • Pattern

Complex object mapping

The conversion from properties to object is using the dotted notation. For instance:

file.path = /home/user/input.csv
file.format = CSV

will match

public class FileOptions {
    private File path;

    private Format format;

assuming the method parameter was configured with @Option("file").

List case

Lists use the same syntax but to define their elements their rely on an indexed syntax. Assuming the list parameter is named files and the elements are of  FileOptions type, here is how to define a list of 2 elements:

files[0].path = /home/user/input1.csv
files[0].format = CSV
files[1].path = /home/user/input2.xml
files[1].format = EXCEL
Map case

Inspired from the list case, the map uses .key[index] and .value[index] to represent its key and values:

// Map<String, FileOptions>
files.key[0] = first-file
files.value[0].path = /home/user/input1.csv
files.value[0].type = CSV
files.key[1] = second-file
files.value[1].path = /home/user/input2.xml
files.value[1].type = EXCEL
// Map<FileOptions, String>
files.key[0].path = /home/user/input1.csv
files.key[0].type = CSV
files.value[0] = first-file
files.key[1].path = /home/user/input2.xml
files.key[1].type = EXCEL
files.value[1] = second-file
don’t abuse of map type. If not needed for your configuration (= if you can configure your component with an object) don’t use it.

Constraints and validation on the configuration/input

It is common to need to add as metadata a field is required, another has a minimum size etc. This is done with the validation in org.talend.sdk.component.api.configuration.constraint package:

API Name Parameter Type Description Supported Types Metadata sample




Ensure the decorated option size is validated with a higher bound.






Ensure the decorated option size is validated with a lower bound.






Validate the decorated string with a javascript pattern (even into the Studio).






Ensure the decorated option size is validated with a higher bound.

Number, int, short, byte, long, double, float





Ensure the decorated option size is validated with a lower bound.

Number, int, short, byte, long, double, float





Mark the field as being mandatory.






Ensure the decorated option size is validated with a higher bound.






Ensure the decorated option size is validated with a lower bound.






Ensure the elements of the collection must be distinct (kind of set).



using the programmatic API the metadata are prefixed by tcomp:: but this prefix is stripped in the web for convenience, the previous table uses the web keys.

Marking a configuration as a particular type of data

It is common to classify the incoming data. You can see it as tagging them in several types. The most common ones are the:

  • datastore: all the data you need to connect to the backend

  • dataset: a datastore coupled with all the data you need to execute an action

API Type Description Metadata sample



Mark a model (complex object) as being a dataset.




Mark a model (complex object) as being a datastore (connection to a backend).


the component family associated with a configuration type (datastore/dataset) is always the one related to the component using that configuration.

Those configuration types can be composed to provide one configuration item. For example a dataset type will often need a datastore type to be provided. and a datastore type (that provides the connection information) will be used to create a dataset type.

Those configuration types will also be used at design time to create shared configuration that can be stored and used at runtime.

For example, we can think about a relational database that support JDBC:

  • A datastore may provide:

    • jdbc url, username, password

  • A dataset may be:

    • datastore (that will provide the connection data to the database)

    • table name, data []

The component server will scan all those configuration types and provide a configuration type index. This index can be used for the integration into the targeted platforms (studio, web applications…​)

The configuration type index is represented as a flat tree that contains all the configuration types represented as nodes and indexed by their ids.

Also, every node can point to other nodes. This relation is represented as an array of edges that provide the childes ids.

For example, a configuration type index for the above example will be:

{nodes: {
             "idForDstore": { datastore:"datastore data", edges:[id:"idForDset"] },
             "idForDset":   { dataset:"dataset data" }

It can be needed to define a binding between properties, a set of annotations allows to do it:

API Name Description Metadata Sample



If the evaluation of the element at the location matches value then the element is considered active, otherwise it is deactivated.




Allows to set multiple visibility conditions on the same property.


Target element location is specified as a relative path to current location using Unix path characters. Configuration class delimiter is /. Parent configuration class is specified by ... Thus ../targetProperty denotes a property, which is located in parent configuration class and has name targetProperty.

using the programmatic API the metadata are prefixed by tcomp:: but this prefix is stripped in the web for convenience, the previous table uses the web keys.

Add hints about the rendering based on configuration/component knowledge

In some case it can be needed to add some metadata about the configuration to let the UI render properly the configuration. A simple example is a password value must be hidden and not a simple clear input box. For these cases - when the component developper wants to influence the UI rendering - you can use a particular set of annotations:

API Description Generated property metadata


Provide a default value the UI can use - only for primitive fields.



Allows to sort a class properties.



Request the rendered to do what it thinks is best.



Advanced layout to place properties by row, this is exclusive with @OptionsOrder.



Allow to configure multiple grid layouts on the same class, qualified with a classifier (name)



Put on a configuration class it notifies the UI an horizontal layout is preferred.



Put on a configuration class it notifies the UI a vertical layout is preferred.



Mark a field as being represented by some code widget (vs textarea for instance).



Mark a field as being a credential. It is typically used to hide the value in the UI.



Mark a List<String> or Map<String, String> field as being represented as the component data selector (field names generally or field names as key and type as value).



Mark a field as being represented by a textarea(multiline text input).


using the programmatic API the metadata are prefixed by tcomp:: but this prefix is stripped in the web for convenience, the previous table uses the web keys.
target support should cover org.talend.core.model.process.EParameterFieldType but we need to ensure web renderers is able to handle the same widgets.



There are also other types of validation similar to @Pattern that you can use :

  • @Min, @Max for numbers.

  • @Unique for collection values

  • @Required for required configuration

Registering components

As seen in the Getting Started, you need an annotation to register your component through family method. Multiple components can use the same family value but the pair family+name MUST be unique for the system.

If you desire (recommended) to share the same component family name instead of repeating yourself in all family methods, you can use @Components annotation on the root package of you component, it will enable you to define the component family and the categories the component belongs to (default is Misc if not set). Here is a sample package-info.java:

@Components(name = "my_component_family", categories = "My Category")
package org.talend.sdk.component.sample;

import org.talend.sdk.component.api.component.Components;

For an existing component it can look like:

@Components(name = "Salesforce", categories = {"Business", "Cloud"})
package org.talend.sdk.component.sample;

import org.talend.sdk.component.api.component.Components;

Components metadata

Components can require a few metadata to be integrated in Talend Studio or Cloud platform. Here is how to provide these information. These metadata are set on the component class and belongs to org.talend.sdk.component.api.component package.

API Description


Set an icon key used to represent the component. Note you can use a custom key with custom() method but it is not guaranteed the icon will be rendered properly.


Set the component version, default to 1.


@PartitionMapper(name = "jaxbInput")
public class JaxbPartitionMapper implements Serializable {
    // ...
Management of configuration versions

If some impacting changes happen on the configuration they can be manage through a migration handler at component level (to enable to support trans-model migration).

The @Version annotation supports a migrationHandler method which will take the implementation migrating the incoming configuration to the current model.

For instance if filepath configuration entry from v1 changed to location in v2 you can remap the value to the right key in your MigrationHandler implementation.

it is recommended to not manage all migrations in the handler but rather split it in services you inject in the migration handler (through constructor):
// full component code structure skipped for brievity, kept only migration part
@Version(value = 3, migrationHandler = MyComponent.Migrations.class)
public class MyComponent {
    // the component code...

    private interface VersionConfigurationHandler {
        Map<String, String> migrate(Map<String, String> incomingData);

    public static class Migrations {
        private final List<VersionConfigurationHandler> handlers;

        // VersionConfigurationHandler implementations are decorated with @Service
        public Migrations(final List<VersionConfigurationHandler> migrations) {
            this.handlers = migrations;
            this.handlers.sort(/*some custom logic*/);

        public Map<String, String> migrate(int incomingVersion, Map<String, String> incomingData) {
            Map<String, String> out = incomingData;
            for (MigrationHandler handler : handlers) {
                out = handler.migrate(out);

What is important in this snippet is not much the way the code is organized but rather the fact you organize your migrations the way which fits the best your component. If migrations are not conflicting no need of something fancy, just apply them all but if you need to apply them in order you need to ensure they are sorted. Said otherwise: don’t see this API as a migration API but as a migration callback and adjust the migration code structure you need behind the MigrationHandler based on your component requirements. The service injection enables you to do so.


@PartitionMapper will obviously mark a partition mapper:

@PartitionMapper(family = "demo", name = "my_mapper")
public class MyMapper {

@Emitter is a shortcut for @PartitionMapper when you don’t support distribution. Said otherwise it will enforce an implicit partition mapper execution with an assessor size of 1 and a split returning itself.

@Emitter(family = "demo", name = "my_input")
public class MyInput {

A method decorated with @Processor will be considered as a producer factory:

@Processor(family = "demo", name = "my_processor")
public class MyProcessor {


In the simplest case you should store messages using ResourceBundle properties file in your component module to use internationalization. The location of the properties file should be in the same package as the related component(s) and is named Messages (ex: org.talend.demo.MyComponent will use org.talend.demo.Messages[locale].properties).

Default components keys

Out of the box components are internationalized using the same location logic for the resource bundle and here is the list of supported keys:

Name Pattern Description


the display name of the family


the display name of a configuration type (dataStore or dataSet)


the display name of the component (used by the GUIs)


the display name of the option.


the display name of the option using it class name.


the display name of the enum_name enum value of the enum enum_simple_class_name.


the placeholder of the option.

Example of configuration for a component named list belonging to the family memory (@Emitter(family = "memory", name = "list")):

memory.list._displayName = Memory List

Configuration class are also translatable using the simple class name in the messages properties file. This useful when you have some common configuration shared within multiple components.

If you have a configuration class like :

public class MyConfig {

    private String host;

    private int port;

You can give it a translatable display name by adding ${simple_class_name}.${property_name}._displayName to Messages.properties under the same package as the config class.

MyConfig.host._displayName = Server Host Name
MyConfig.host._placeholder = Enter Server Host Name...

MyConfig.port._displayName = Server Port
MyConfig.port._placeholder = Enter Server Port...
If you have a display name using the property path, it will override the display name defined using the simple class name. this rule apply also to placeholders

Components Packaging

Component Loading

Talend Component scanning is based on a plugin concept. To ensure plugins can be developped in parallel and avoid conflicts it requires to isolate plugins (components or component grouped in a single jar/plugin).

Here we have multiple options which are (high level):

  • flat classpath: listed for completeness but rejected by design because it doesn’t match at all this requirement.

  • tree classloading: a shared classloader inherited by plugin classloaders but plugin classloader classes are not seen by the shared classloader nor by other plugins.

  • graph classloading: this one allows you to link the plugins and dependencies together dynamically in any direction.

If you want to map it to concrete common examples, the tree classloading is commonly used by Servlet containers where plugins are web applications and the graph classloading can be illustrated by OSGi containers.

In the spirit of avoiding a lot of complexity added by this layer, Talend Component relies on a tree classloading. The advantage is you don’t need to define the relationship with other plugins/dependencies (it is built-in).

Here is a representation of this solution:

classloader layout

The interesting part is the shared area will contain Talend Component API which is the only (by default) shared classes accross the whole plugins.

Then each plugins will be loaded in their own classloader with their dependencies.

Packaging a plugin

this part explains the overall way to handle dependecnies but the Talend Maven plugin provides a shortcut for that.

A plugin is just a jar which was enriched with the list of its dependencies. By default Talend Component runtime is able to read the output of maven-dependency-plugin in TALEND-INF/dependencies.txt location so you just need to ensure your component defines the following plugin:


If you check your jar once built you will see that the file contains something like:

$ unzip -p target/mycomponent-1.0.0-SNAPSHOT.jar TALEND-INF/dependencies.txt

The following files have been resolved:

What is important to see is the scope associated to the artifacts:

  • the API (component-api and geronimo-annotation_1.3_spec) are provided because you can consider them to be there when executing (it comes with the framework)

  • your specific dependencies (awesome-project) is compile: it will be included as a needed dependency by the framework (note that using runtime works too).

  • the other dependencies will be ignored (test dependencies)

Packaging an application

Even if a flat classpath deployment is possible, it is not recommended because it would then reduce the capabilities of the components.


The way the framework resolves dependencies is based on a local maven repository layout. As a quick reminder it looks like:

├── groupId1
│   └── artifactId1
│       ├── version1
│       │   └── artifactId1-version1.jar
│       └── version2
│           └── artifactId1-version2.jar
└── groupId2
    └── artifactId2
        └── version1
            └── artifactId2-version1.jar

This is all the layout the framework will use. Concretely the logic will convert the t-uple {groupId, artifactId, version, type (jar)} to the path in the repository.

Talend Component runtime has two ways to find an artifact:

  • from the file system based on a configure maven 2 repository.

  • from a fatjar (uber jar) with a nested maven repository under MAVEN-INF/repository.

The first option will use either - by default - ${user.home}/.m2/repository or a specific path configured when creating a ComponentManager. The nested repository option will need some configuration during the packaging to ensure the repository is well created.

Create a nested maven repository with maven-shade-plugin

To create the nested MAVEN-INF/repository repository you can use nested-maven-repository extension:

          <transformer implementation="org.talend.sdk.component.container.maven.shade.ContainerDependenciesTransformer">
Listing needed plugins

Plugin are programmatically registered in general but if you want to make some of them automatically available you need to generate a TALEND-INF/plugins.properties which will map a plugin name to coordinates found with the maven mecanism we just talked about.

Here again we can enrich maven-shade-plugin to do it:

          <transformer implementation="org.talend.sdk.component.container.maven.shade.PluginTransformer">
maven-shade-plugin extensions

Here is a final job/application bundle based on maven shade plugin:

          <transformer implementation="org.talend.sdk.component.container.maven.shade.PluginTransformer">
the configuration unrelated to transformers can depend your application.

ContainerDependenciesTransformer is the one to embed a maven repository and PluginTransformer to create a file listing (one per line) a list of artifacts (representing plugins).

Both transformers share most of their configuration:

  • session: must be set to ${session}. This is used to retrieve dependencies.

  • scope: a comma separated list of scope to include in the artifact filtering (note that the default will rely on provided but you can replace it by compile, runtime, runtime+compile, runtime+system, test).

  • include: a comma separated list of artifact to include in the artifact filtering.

  • exclude: a comma separated list of artifact to exclude in the artifact filtering.

  • userArtifacts: a list of artifacts (groupId, artifactId, version, type - optional, file - optional for plugin transformer, scope - optional) which can be forced inline - mainly useful for PluginTransformer.

  • includeTransitiveDependencies: should transitive dependencies of the components be included, true by default.

  • includeProjectComponentDependencies: should project component dependencies be included, false by default (normally a job project uses isolation for components so this is not needed).

  • userArtifacts: set of component artifacts to include.

to use with the component tooling, it is recommended to keep default locations. Also if you feel you need to use project dependencies, you can need to refactor your project structure to ensure you keep component isolation. Talend component let you handle that part but the recommended practise is to use userArtifacts for the components and not the project <dependencies>.

ContainerDependenciesTransformer specific configuration is the following one:

  • repositoryBase: base repository location (default to MAVEN-INF/repository).

  • ignoredPaths: a comma separated list of folder to not create in the output jar, this is common for the ones already created by other transformers/build parts.


ContainerDependenciesTransformer specific configuration is the following one:

  • pluginListResource: base repository location (default to TALEND-INF/plugins.properties`).

Example: if you want to list only the plugins you use you can configure this transformer like that:

<transformer implementation="org.talend.sdk.component.container.maven.shade.PluginTransformer">

Component scanning rules and default exclusions

The framework uses two kind of filtering when scanning your component. One based on the jar name and one based on the package name. Ensure your component definitions (including services) are in a scanned module if not registered manually using ComponentManager.instance().addPlugin() and that its package is not excluded.

Jars Scanning

To find components the framework can scan the classpath but in this case, to avoid to scan the whole classpath which can be really huge an impacts a lot the startup time, several jars are excluded out of the box.

These jars use the following prefix:

  • ApacheJMeter

  • FastInfoset

  • HdrHistogram

  • HikariCP

  • PDFBox

  • RoaringBitmap-

  • XmlSchema-

  • accessors-smart

  • activation-

  • activeio-

  • activemq-

  • aeron

  • aether-

  • agrona

  • akka-

  • animal-sniffer-annotation

  • annotation

  • ant-

  • antlr-

  • aopalliance-

  • apache-el

  • apacheds-

  • api-asn1-

  • api-util-

  • apiguardian-api-

  • app-

  • archaius-core

  • args4j-

  • arquillian-

  • asciidoctorj-

  • asm-

  • aspectj

  • async-http-client-

  • avalon-framework-

  • avro-

  • awaitility-

  • axis-

  • axis2-

  • base64-

  • batchee-jbatch

  • batik-

  • bcpkix

  • bcprov-

  • beam-model-

  • beam-runners-

  • beam-sdks-

  • bonecp

  • bootstrap.jar

  • brave-

  • bsf-

  • build-link

  • bval

  • byte-buddy

  • c3p0-

  • cache

  • carrier

  • cassandra-driver-core

  • catalina-

  • catalina.jar

  • cats

  • cdi-

  • cglib-

  • charsets.jar

  • chill

  • classindex

  • classmate

  • classutil

  • classycle

  • cldrdata

  • commands-

  • common-

  • commons-

  • component-api

  • component-form

  • component-runtime

  • component-server

  • component-spi

  • component-studio

  • components-api

  • components-common

  • compress-lzf

  • config

  • constructr

  • container-core

  • contenttype

  • coverage-agent

  • cryptacular-

  • cssparser-

  • curator-

  • cxf-

  • daikon

  • databinding

  • dataquality

  • debugger-agent

  • deltaspike-

  • deploy.jar

  • derby-

  • derbyclient-

  • derbynet-

  • dnsns

  • dom4j

  • draw2d

  • ecj-

  • eclipselink-

  • ehcache-

  • el-api

  • enunciate-core-annotations

  • error_prone_annotations

  • expressions

  • fastutil

  • feign-core

  • feign-hystrix

  • feign-slf4j

  • filters-helpers

  • findbugs-

  • fluentlenium-core

  • freemarker-

  • fusemq-leveldb-

  • gef-

  • geocoder

  • geronimo-

  • gmbal

  • google-

  • gpars-

  • gragent.jar

  • graph

  • grizzled-scala

  • grizzly-

  • groovy-

  • grpc-

  • gson-

  • guava-

  • guice-

  • h2-

  • hadoop-

  • hamcrest-

  • hawtbuf-

  • hawtdispatch-

  • hawtio-

  • hawtjni-runtime

  • help-

  • hibernate-

  • hk2-

  • howl-

  • hsqldb-

  • htmlunit-

  • htrace-

  • httpclient-

  • httpcore-

  • httpmime

  • hystrix

  • iban4j-

  • icu4j-

  • idb-

  • idea_rt.jar

  • instrumentation-api

  • istack-commons-runtime-

  • ivy-

  • j2objc-annotations

  • jBCrypt

  • jaccess

  • jackson-

  • janino-

  • jansi-

  • jasper-el.jar

  • jasper.jar

  • jasypt-

  • java-atk-wrapper

  • java-support-

  • java-xmlbuilder-

  • javacsv

  • javaee-

  • javaee-api

  • javassist-

  • javaws.jar

  • javax.

  • jaxb-

  • jaxp-

  • jbake-

  • jboss-

  • jbossall-

  • jbosscx-

  • jbossjts-

  • jbosssx-

  • jcache

  • jce.jar

  • jcip-annotations

  • jcl-over-slf4j-

  • jcommander-

  • jdbcdslog-1

  • jersey-

  • jets3t

  • jettison-

  • jetty-

  • jface

  • jfairy

  • jffi

  • jfr.jar

  • jfxrt.jar

  • jfxswt

  • jjwt

  • jline

  • jmdns-

  • jmustache

  • jna-

  • jnr-

  • jobs-

  • joda-convert

  • joda-time-

  • johnzon-

  • jolokia-

  • jopt-simple

  • jruby-

  • json-

  • json4s-

  • jsonb-api

  • jsoup-

  • jsp-api

  • jsr

  • jsse.jar

  • jta

  • jul-to-slf4j-

  • juli-

  • junit-

  • junit5-

  • jwt

  • kafka

  • kahadb-

  • kotlin-runtime

  • kryo

  • leveldb

  • libphonenumber

  • lift-json

  • lmdbjava

  • localedata

  • log4j-

  • logback

  • logging-event-layout

  • logkit-

  • lombok

  • lucene

  • lz4

  • machinist

  • macro-compat

  • mail-

  • management-

  • mapstruct-

  • maven-

  • mbean-annotation-api-

  • meecrowave-

  • mesos-

  • metrics-

  • mimepull-

  • mina-

  • minlog

  • mockito-core

  • mqtt-client-

  • multiverse-core-

  • mx4j-

  • myfaces-

  • mysql-connector-java-

  • nashorn

  • neethi-

  • neko-htmlunit

  • nekohtml-

  • netflix

  • netty-

  • nimbus-jose-jwt

  • objenesis-

  • okhttp

  • okio

  • openjpa-

  • openmdx-

  • opensaml-

  • opentest4j-

  • openwebbeans-

  • openws-

  • ops4j-

  • org.apache.aries

  • org.apache.commons

  • org.apache.log4j

  • org.eclipse.

  • org.junit.

  • org.osgi.core-

  • org.osgi.enterprise

  • org.talend

  • orient-commons-

  • orientdb-core-

  • orientdb-nativeos-

  • oro-

  • osgi

  • paranamer

  • pax-url

  • play

  • plexus-

  • plugin.jar

  • poi-

  • postgresql

  • preferences-

  • prefixmapper

  • protobuf-

  • py4j-

  • pyrolite-

  • qdox-

  • quartz-2

  • quartz-openejb-

  • reactive-streams

  • reflectasm-

  • registry-

  • resources.jar

  • ribbon

  • rmock-

  • routes-compiler

  • routines

  • rt.jar

  • runners

  • runtime-

  • rxjava

  • rxnetty

  • saaj-

  • sac-

  • scala

  • scalap

  • scalatest

  • scannotation-

  • selenium

  • serializer-

  • serp-

  • service-common

  • servlet-api-

  • servo-

  • shaded

  • shrinkwrap-

  • sisu-guice

  • sisu-inject

  • slf4j-

  • slick

  • smack-

  • smackx-

  • snakeyaml-

  • snappy-

  • spark-

  • specs2

  • spring-

  • sshd-

  • ssl-config-core

  • stax-api-

  • stax2-api-

  • stream

  • sunec.jar

  • sunjce_provider

  • sunpkcs11

  • surefire-

  • swagger-

  • swizzle-

  • sxc-

  • system-rules

  • tachyon-

  • talend-icon

  • test-interface

  • testng-

  • tomcat

  • tomee-

  • tools.jar

  • twirl

  • twitter4j-

  • tyrex

  • uncommons

  • unused

  • validation-api-

  • velocity-

  • wagon-

  • webbeans-

  • websocket

  • woodstox-core

  • workbench

  • ws-commons-util-

  • wsdl4j-

  • wss4j-

  • wstx-asl-

  • xalan-

  • xbean-

  • xercesImpl-

  • xml-apis-

  • xml-resolver-

  • xmlbeans-

  • xmlenc-

  • xmlgraphics-

  • xmlpull-

  • xmlrpc-

  • xmlschema-

  • xmlsec-

  • xmltooling-

  • xmlunit-

  • xstream-

  • xz-

  • zipfs.jar

  • zipkin-

  • ziplock-

  • zkclient

  • zookeeper-

Package Scanning

Since the framework can be used in the case of fatjars or shades, and because it still uses scanning, it is important to ensure we don’t scan the whole classes for performances reason.

Therefore, the following packages are ignored:

  • avro.shaded

  • com.codehale.metrics

  • com.ctc.wstx

  • com.datastax.driver.core

  • com.fasterxml.jackson.annotation

  • com.fasterxml.jackson.core

  • com.fasterxml.jackson.databind

  • com.fasterxml.jackson.dataformat

  • com.fasterxml.jackson.module

  • com.google.common

  • com.google.thirdparty

  • com.ibm.wsdl

  • com.jcraft.jsch

  • com.kenai.jffi

  • com.kenai.jnr

  • com.sun.istack

  • com.sun.xml.bind

  • com.sun.xml.messaging.saaj

  • com.sun.xml.txw2

  • com.thoughtworks

  • io.jsonwebtoken

  • io.netty

  • io.swagger.annotations

  • io.swagger.config

  • io.swagger.converter

  • io.swagger.core

  • io.swagger.jackson

  • io.swagger.jaxrs

  • io.swagger.model

  • io.swagger.models

  • io.swagger.util

  • javax

  • jnr

  • junit

  • net.sf.ehcache

  • net.shibboleth.utilities.java.support

  • org.aeonbits.owner

  • org.apache.activemq

  • org.apache.beam

  • org.apache.bval

  • org.apache.camel

  • org.apache.catalina

  • org.apache.commons.beanutils

  • org.apache.commons.cli

  • org.apache.commons.codec

  • org.apache.commons.collections

  • org.apache.commons.compress

  • org.apache.commons.dbcp2

  • org.apache.commons.digester

  • org.apache.commons.io

  • org.apache.commons.jcs.access

  • org.apache.commons.jcs.admin

  • org.apache.commons.jcs.auxiliary

  • org.apache.commons.jcs.engine

  • org.apache.commons.jcs.io

  • org.apache.commons.jcs.utils

  • org.apache.commons.lang

  • org.apache.commons.lang3

  • org.apache.commons.logging

  • org.apache.commons.pool2

  • org.apache.coyote

  • org.apache.cxf

  • org.apache.geronimo.javamail

  • org.apache.geronimo.mail

  • org.apache.geronimo.osgi

  • org.apache.geronimo.specs

  • org.apache.http

  • org.apache.jcp

  • org.apache.johnzon

  • org.apache.juli

  • org.apache.logging.log4j.core

  • org.apache.logging.log4j.jul

  • org.apache.logging.log4j.util

  • org.apache.logging.slf4j

  • org.apache.meecrowave

  • org.apache.myfaces

  • org.apache.naming

  • org.apache.neethi

  • org.apache.openejb

  • org.apache.openjpa

  • org.apache.oro

  • org.apache.tomcat

  • org.apache.tomee

  • org.apache.velocity

  • org.apache.webbeans

  • org.apache.ws

  • org.apache.wss4j

  • org.apache.xbean

  • org.apache.xml

  • org.apache.xml.resolver

  • org.bouncycastle

  • org.codehaus.jackson

  • org.codehaus.stax2

  • org.codehaus.swizzle.Grep

  • org.codehaus.swizzle.Lexer

  • org.cryptacular

  • org.eclipse.jdt.core

  • org.eclipse.jdt.internal

  • org.fusesource.hawtbuf

  • org.h2

  • org.hamcrest

  • org.hsqldb

  • org.jasypt

  • org.jboss.marshalling

  • org.joda.time

  • org.jose4j

  • org.junit

  • org.jvnet.mimepull

  • org.metatype.sxc

  • org.objectweb.asm

  • org.objectweb.howl

  • org.openejb

  • org.opensaml

  • org.slf4j

  • org.swizzle

  • org.terracotta.context

  • org.terracotta.entity

  • org.terracotta.modules.ehcache

  • org.terracotta.statistics

  • org.tukaani

  • org.yaml.snakeyaml

  • serp

it is not recommanded but possible to add in your plugin module a TALEND-INF/scanning.properties file with classloader.includes and classloader.excludes entries to refine the scanning with custom rules. In such a case, exclusions win over inclusions.

Build tools

Maven Plugin

talend-component-maven-plugin intends to help you to write components validating components match best practices and also generating transparently metadata used by Talend Studio.

Here is how to use it:


Note that this plugin is also an extension so you can declare it in your build/extensions block as:


Used as an extension, dependencies, validate and documentation goals will be set up.


The first goal is a shortcut for the maven-dependency-plugin, it will create the TALEND-INF/dependencies.txt file with the compile and runtime dependencies to let the component use it at runtime:



The most important goal is here to help you to validate the common programming model of the component. Here is the execution definition to activate it:


By default it will be bound to process-classes phase. When executing it will do several validations which can be switched off adding the corresponding flags to false in the <configuration> block of the execution:

Name Description Default


Validates resource bundle are presents and contain commonly used keys (like _displayName)



Ensure components pass validations of the ComponentManager and Talend Component runtime



Ensure components are Serializable - note this is a sanity check, the component is not actually serialized here, if you have a doubt ensure to test it. It also checks any @Internationalized class is valid and has its keys.



Ensure components define an @Icon and @Version.



Ensure any @DataStore defines a @HealthCheck.



Ensure native programming model is respected, you can disable it when using another programming model like in beam case.



Validate actions signatures for the ones not tolerating dynamic binding (@HealthCheck, @DynamicValues, …​). It is recommended to keep it true.



Validate the family, i.e. the package containing the @Components has also a @Icon.



Ensure all 1. components and 2. @Option properties have a documentation using @Documentation



This goal generates an Asciidoc file documenting your component from the configuration model (@Option) and @Documentation you can put on options and the component itself.

Name Description Default


Which level are the root title

2 which means ==


Where to store the output, it is NOT recommended to change it



A map of the renderings to do, keys are the format (pdf or html) and values the output paths



A map of asciidoctor attributes when formats is set


templateDir / templateEngine

Template configuration for the rendering



Document title



Should the documentations (.adoc, and formats keys) should be attached to the project (and deployed)


if you use the extension you can add the property talend.documentation.htmlAndPdf and set it to true in your project to automatically get a html and PDF rendering of the documentation.
Render your documentation

To render the generated documentation you can use the Asciidoctor Maven plugin (or Gradle equivalent):

<plugin> (1)
<plugin> (2)
  1. Will generate in target/classes/TALEND-INF/documentation.adoc the components documentation.

  2. Will render the documenation as an html file in target/documentation/documentation.html.

ensure to execute it after the documentation generation.

If you prefer a PDF rendering you can configure the following execution in the asciidoctor plugin (note that you can configure both executions if you want both HTML and PDF rendering):

Include the documentation into a document

If you want to add some more content or add a title, you can include the generated document into another document using Asciidoc include directive.

A common example is:

= Super Components
Super Writer
:toclevels: 3
:source-highlighter: prettify
:icons: font
:imagesdir: images


This assumes you pass to the plugin the attribute generated_doc, this can be done this way:


This is optional but allows to reuse maven placeholders to pass paths which is quite convenient in an automated build.


You can find more customizations on Asciidoctor website.


Testing the rendering of your component(s) configuration into the Studio is just a matter of deploying a component in Talend Studio (you can have a look to link::studio.html[Studio Documentation] page. But don’t forget the component can also be deployed into a Cloud (web) environment. To ease the testing of the related rendering, you can use the goal web of the plugin:

mvn talend-component:web

Then you can test your component going on localhost:8080. You need to select which component form you want to see using the treeview on the left, then on the right the form will be displayed.

The two available configurations of the plugin are serverPort which is a shortcut to change the default, 8080, port of the embedded server and serverArguments to pass Meecrowave options to the server. More on that configuration is available at openwebbeans.apache.org/meecrowave/meecrowave-core/cli.html.

this command reads the component jar from the local maven repository so ensure to install the artifact before using it.

Generate inputs or outputs

The Mojo generate (maven plugin goal) of the same plugin also embeds a generator you can use to bootstrap any input or output component:

    <execution> (1)
    <execution> (2)
1 Generates an input (partition mapper + emitter)
2 Generates an output

It is intended to be used from the command line (or IDE Maven integration):

$ mvn talend-component:generate \
    -Dtalend.generator.type=[input|output] \ (1)
    [-Dtalend.generator.classbase=com.test.MyComponent] \ (2)
    [-Dtalend.generator.family=my-family] \ (3)
    [-Dtalend.generator.pom.read-only=false] (4)
1 select the type of component you want, input to generate a mapper and emitter and output to generate an output processor
2 set the class name base (will be suffixed by the component type), if not set the package will be guessed and classname based on the basedir name
3 set the component family to use, default to the base dir name removing (component[s] from the name, ex: my-component will lead to my as family if not explicitly set)
4 should the generator try to add component-api in the pom if not already here, if you added it you can set it to false directly in the pom

For this command to work you will need to just register the plugin:


Talend Component Archive

Component ARchive (.car) is the way to bundle a component to share it in Talend ecosystem. It is a plain Java ARchive (.jar) containing a metadata file and a nested maven repository containing the component and its depenencies.

mvn talend-component:car

It will create a .car in your build directory which is shareable on Talend platforms.

Note that this CAR is executable and exposes the command studio-deploy which takes as parameter a Talend Studio home location. Executed it will install the dependencies into the studio and register the component in your instance. Here is a sample launch command:

# for a studio
java -jar mycomponent.car studio-deploy /path/to/my/studio

# for a m2 provisioning
java -jar mycomponent.car maven-deploy /path/to/.m2/repository

Gradle Plugin

gradle-talend-component intends to help you to write components validating components match best practices. It is inspired from the Maven plugin and adds the ability to generate automatically the dependencies.txt file the SDK uses to build the component classpath. For more information on the configuration you can check out the maven properties matching the attributes.

Here is how to use it:

buildscript {
  repositories {
  dependencies {
    classpath "org.talend.sdk.component:gradle-talend-component:${talendComponentVersion}"

apply plugin: 'org.talend.sdk.component'
apply plugin: 'java'

// optional customization
talendComponentKit {
    // dependencies.txt generation, replaces maven-dependency-plugin
    dependenciesLocation = "TALEND-INF/dependencies.txt"
    boolean skipDependenciesFile = false;

    // classpath for validation utilities
    sdkVersion = "${talendComponentVersion}"
    apiVersion = "${talendComponentApiVersion}"

    // documentation
    skipDocumentation = false
    documentationOutput = new File(....)
    documentationLevel = 2 // first level will be == in the generated adoc
    documentationTitle = 'My Component Family' // default to project name
    documentationFormats = [:] // adoc attributes
    documentationFormats = [:] // renderings to do

    // validation
    skipValidation = false
    validateFamily = true
    validateSerializable = true
    validateInternationalization = true
    validateModel = true
    validateMetadata = true
    validateComponent = true
    validateDataStore = true
    validateDataSet = true
    validateActions = true

    // web
    serverArguments = []
    serverPort = 8080

    // car
    carOutput = new File(....)
    carMetadata = [:] // custom meta (string key-value pairs)



Recommanded practise for internationalization are:

  • store messages using ResourceBundle properties file in your component module

  • the location of the properties are in the same package than the related component(s) and is named Messages (ex: org.talend.demo.MyComponent will use org.talend.demo.Messages[locale].properties)

  • for your own messages use the internationalization API

Internationalization API

Overal idea is to design its messages as methods returning String values and back the template by a ResourceBundle located in the same package than the interface defining these methods and named Messages.

this is the mecanism to use to internationalize your own messages in your own components.

To ensure you internationalization API is identified you need to mark it with @Internationalized:

@Internationalized (1)
public interface Translator {

    String message();

    String templatizedMessage(String arg0, int arg1); (2)

    String localized(String arg0, @Language Locale locale); (3)
1 @Internationalized allows to mark a class as a i18n service
2 you can pass parameters and the message will use MessageFormat syntax to be resolved based on the ResourceBundle template
3 you can use @Language on a Locale parameter to specify manually the locale to use, note that a single value will be used (the first parameter tagged as such).

Providing some actions for consumers/clients

In some cases you will desire to add some actions unrelated to the runtime. A simple example is to enable clients - the users of the plugin/library - to test if a connection works. Even more concretely: does my database is up?.

To do so you need to define an @Action which is a method with a name (representing the event name) in a class decorated with @Service:

public class MyDbTester {
    @Action(family = "mycomp", "test")
    public Status doTest(final IncomingData data) {
        return ...;
services are singleton so if you need some thread safety ensure they match that requirement. They shouldn’t store any state too (state is held by the component) since they can be serialized any time.
services are usable in components as well (matched by type) and allow to reuse some shared logic like a client. Here is a sample with a service used to access files:
@Emitter(family = "sample", name = "reader")
public class PersonReader implements Serializable {
    // attributes skipped to be concise

    public PersonReader(@Option("file") final File file,
                        final FileService service) {
        this.file = file;
        this.service = service;

    // use the service
    public void open() throws FileNotFoundException {
        reader = service.createInput(file);

service is passed to constructor automatically, it can be used as a bean. Only call of service’s method is required.

Particular action types

Some actions are that common and need a clear contract so they are defined as API first citizen, this is the case for wizards or healthchecks for instance. Here is the list of all actions:

API Type Description Return type Sample returned type



Mark a method as being useful to fill potential values of a string option for a property denoted by its value. You can link a field as being completable using @Proposable(value). The resolution of the completion action is then done through the component family and value of the action. The callback doesn’t take any parameter.





This class marks an action doing a connection test


{"comment":"Something went wrong","status":"KO"}



Mark an action as returning a discovered schema. Its parameter MUST be the type decorated with @Structure.










Mark a method as being used to validate a configuration. IMPORTANT: this is a server validation so only use it if you can’t use other client side validation to implement it.


{"comment":"Something went wrong","status":"KO"}

Built in services

The framework provides some built-in services you can inject by type in components and actions out of the box.

Here is the list:

Type Description


Provides a small abstraction to cache data which don’t need to be recomputed very often. Commonly used by actions for the UI interactions.


Allows to resolve a dependency from its Maven coordinates.


A JSON-B instance. If your model is static and you don’t want to handle the serialization manually using JSON-P you can inject that instance.


A JSON-P instance. Prefer other JSON-P instances if you don’t exactly know why you use this one.


A JSON-P instance. It is recommended to use this one instead of a custom one for memory/speed optimizations.


A JSON-P instance. It is recommended to use this one instead of a custom one for memory/speed optimizations.


A JSON-P instance. It is recommended to use this one instead of a custom one for memory/speed optimizations.


A JSON-P instance. It is recommended to use this one instead of a custom one for memory/speed optimizations.


A JSON-P instance. It is recommended to use this one instead of a custom one for memory/speed optimizations.


Represents the local configuration which can be used during the design.

it is not recommended to use it for the runtime since the local configuration is generally different and the instances are distincts.
you can also use the local cache as an interceptor with @Cached

Every interface that extends HttpClient and that contains methods annotated with @Request

This let you define an http client in a declarative manner using an annotated interface.

See the HttpClient usage for details.
all these injected instances are serializable which is important for the big data environment, if you create the instances yourself you will not benefit from that features and the memory optimization done by the runtime so try to prefer to reuse the framework instances over custom ones.

HttpClient usage

Let assume that we have a REST API defined like below, and that it requires a basic authentication header.

GET /api/records/{id}


POST /api/records

with a json playload to be created {"id":"some id", "data":"some data"}

To create an http client able to consume this REST API, we will define an interface that extends HttpClient,

The HttpClient interface lets you set the base for the http address that our client will hit.

The base is the part of the address that we will need to add to the request path to hit the api.

Every method annotated with @Request of our interface will define an http request. Also every request can have @Codec that let us encode/decode the request/response playloads.

if your payload(s) is(are) String or Void you can ignore the coder/decoder.
public interface APIClient extends HttpClient {
    @Request(path = "api/records/{id}", method = "GET")
    @Codec(decoder = RecordDecoder.class) //decoder =  decode returned data to Record class
    Record getRecord(@Header("Authorization") String basicAuth, @Path("id") int id);

    @Request(path = "api/records", method = "POST")
    @Codec(encoder = RecordEncoder.class, decoder = RecordDecoder.class) //encoder = encode record to fit request format (json in this example)
    Record createRecord(@Header("Authorization") String basicAuth, Record record);
The interface should extends HttpClient.

In the codec classes (class that implement Encoder/Decoder) you can inject any of your services annotated with @Service or @Internationalized into the constructor. The i18n services can be useful to have i18n messages for errors handling for example.

This interface can be injected into our Components classes or Services to consume the defined api.

public class MyService {

    private APIClient client;

    public MyService(...,APIClient client){
        this.client = client;
        client.base("http://localhost:8080");// init the base of the api, ofen in a PostConstruct or init method

    // Our get request
    Record rec =  client.getRecord("Basic MLFKG?VKFJ", 100);

    // Our post request
    Record newRecord = client.createRecord("Basic MLFKG?VKFJ", new Record());

Note: by default /+json are mapped to JSON-P and /+xml to JAX-B if the model has a @XmlRootElement annotation.

Advanced HTTP client request customization

For advanced cases you can customize the Connection directly using @UseConfigurer on the method. It will call your custom instance of Configurer. Note that you can use some @ConfigurerOption in the method signature to pass some configurer configuration.

For instance if you have this configurer:

public class BasicConfigurer implements Configurer {
    public void configure(final Connection connection, final ConfigurerConfiguration configuration) {
        final String user = configuration.get("username", String.class);
        final String pwd = configuration.get("password", String.class);
            Base64.getEncoder().encodeToString((user + ':' + pwd).getBytes(StandardCharsets.UTF_8)));

You can then set it on a method to automatically add the basic header with this kind of API usage:

public interface APIClient extends HttpClient {
    @Request(path = "...")
    Record findRecord(@ConfigurerOption("username") String user, @ConfigurerOption("password") String pwd);

Services and interceptors

For common concerns like caching, auditing etc, it can be fancy to use interceptor like API. It is enabled by the framework on services.

An interceptor defines an annotation marked with @Intercepts which defines the implementation of the interceptor (an InterceptorHandler).

Here is an example:

@Target({ TYPE, METHOD })
public @interface Logged {
    String value();

Then handler is created from its constructor and can take service injections (by type). The first parameter, however, can be a BiFunction<Method, Object[], Object> which representes the invocation chain if your interceptor can be used with others.

if you do a generic interceptor it is important to pass the invoker as first parameter. If you don’t do so you can’t combine interceptors at all.

Here is an interceptor implementation for our @Logged API:

public class LoggingHandler implements InterceptorHandler {
    // injected
    private final BiFunction<Method, Object[], Object> invoker;
    private final SomeService service;

    // internal
    private final ConcurrentMap<Method, String> loggerNames = new ConcurrentHashMap<>();

    public CacheHandler(final BiFunction<Method, Object[], Object> invoker, final SomeService service) {
        this.invoker = invoker;
        this.service = service;

    public Object invoke(final Method method, final Object[] args) {
        final String name = loggerNames.computeIfAbsent(method, m -> findAnnotation(m, Logged.class).get().value());
        service.getLogger(name).info("Invoking {}", method.getName());
        return invoker.apply(method, args);

This implementation is compatible with interceptor chains since it takes the invoker as first constructor parameter and it also takes a service injection. Then the implementation just does what is needed - logging the invoked method here.

the findAnnotation annotation - inherited from InterceptorHandler is an utility method to find an annotation on a method or class (in this order).

Creating a job pipeline

Job Builder

The Job builder let you create a job pipeline programmatically using Talend components (Producers and Processors). The job pipeline is an acyclic graph, so you can built complex pipelines.

Let’s take a simple use case where we will have 2 data source (employee and salary) that we will format to csv and write the result to a file.

A job is defined based on components (nodes) and links (edges) to connect their branches together.

Every component is defined by an unique id and an URI that identify the component.

The URI follow the form : [family]://[component][?version][&configuration]

  • family: the name of the component family

  • component: the name of the component

  • version : the version of the component, it’s represented in a key=value format. where the key is __version and the value is a number.

  • configuration: here you can provide the component configuration as key=value tuple where the key is the path of the configuration and the value is the configuration value in string format.

URI Example
configuration parameters must be URI/URL encoded.

Here is a more concrete job example:

Job.components()   (1)
        .component("salary", "db://input")
        .component("concat", "transform://concat?separator=;")
        .component("csv", "file://out?__version=2")
    .connections()  (2)
        .from("employee").to("concat", "string1")
        .from("salary").to("concat", "string2")
    .build()    (3)
    .run(); (4)
1 We define all the components that will be used in the job pipeline.
2 Then, we define the connections between the components to construct the job pipeline. the links fromto use the component id and the default input/output branches. You can also connect a specific branch of a component if it has multiple or named inputs/outputs branches using the methods from(id, branchName)to(id, branchName). In the example above, the concat component have to inputs (string1 and string2).
3 In this step, we validate the job pipeline by asserting that :
  • It has some starting components (component that don’t have a from connection and that need to be of type producer).

  • There is no cyclic connections. as the job pipeline need to be an acyclic graph.

  • All the components used in connections are already declared.

  • The connection is used only once. you can’t connect a component input/output branch twice.

4 We run the job pipeline.
In this version, the execution of the job is linear. the component are not executed in parallel even if some steps may be independents.

Depending the configuration you can select which environment you execute your job in.

To select the environment the logic is the following one:

  1. if an org.talend.sdk.component.runtime.manager.chain.Job.ExecutorBuilder is passed through the job properties then use it (supported type are a ExecutionBuilder instance, a Class or a String).

  2. if an ExecutionBuilder SPI is present then use it (it is the case if component-runtime-beam is present in your classpath).

  3. else just use a local/standalone execution.

In the case of a Beam execution you can customize the pipeline options using system properties. They have to be prefixed by talend.beam.job.. For instance to set appName option you will set -Dtalend.beam.job.appName=mytest.

Key Provider

The job builder let you set a key provider to join your data when a component has multiple inputs. The key provider can be set contextually to a component or globally to the job

                 (GroupKeyProvider) context -> context.getData().getString("id")) (1)
        .component("salary", "db://input")
        .component("concat", "transform://concat?separator=;")
        .from("employee").to("concat", "string1")
        .from("salary").to("concat", "string2")
    .property(GroupKeyProvider.class.getName(), (2)
                 (GroupKeyProvider) context -> context.getData().getString("employee_id"))
1 Here we have defined a key provider for the data produced by the component employee
2 Here we have defined a key provider for all the data manipulated in this job.

If the incoming data has different ids you can provide a complex global key provider relaying on the context that give you the component id and the branch Name.

GroupKeyProvider keyProvider = context -> {
    if ("employee".equals(context.getComponentId())) {
        return context.getData().getString("id");
    return context.getData().getString("employee_id");

Beam case

For beam case, you need to rely on beam pipeline definition and use component-runtime-beam dependency which provides Beam bridges.


org.talend.sdk.component.runtime.beam.TalendIO provides a way to convert a partition mapper or a processor to an input or processor using the read or write methods.

public class Main {
    public static void main(final String[] args) {
        final ComponentManager manager = ComponentManager.instance()
        Pipeline pipeline = Pipeline.create();
        //Create beam input from mapper and apply input to pipeline
        pipeline.apply(TalendIO.read(manager.findMapper(manager.findMapper("sample", "reader", 1, new HashMap<String, String>() {{
                    put("fileprefix", "input");
                .apply(new ViewsMappingTransform(emptyMap(), "sample")) // prepare it for the output record format (see next part)
        //Create beam processor from talend processor and apply to pipeline
                .apply(TalendIO.write(manager.findProcessor("test", "writer", 1, new HashMap<String, String>() {{
                    put("fileprefix", "output");
                }}).get(), emptyMap()));

        //... run pipeline

org.talend.sdk.component.runtime.beam.TalendFn provides the way to wrap a processor in a Beam PTransform and integrate it in the pipeline.

public class Main {
    public static void main(final String[] args) {
        //Component manager and pipeline initialization...

        //Create beam PTransform from processor and apply input to pipeline
        pipeline.apply(TalendFn.asFn(manager.findProcessor("sample", "mapper", 1, emptyMap())).get())), emptyMap());

        //... run pipeline

The multiple inputs/outputs are represented by a Map element in beam case to avoid to use multiple inputs/outputs.

you can use ViewsMappingTransform or CoGroupByKeyResultMappingTransform to adapt the input/output format to the record format representing the multiple inputs/output, so a kind of Map<String, List<?>>, but materialized as a JsonObject. Input data must be of type JsonObject in this case.
Convert a Beam.io in a component I/O

For simple I/O you can get automatic conversion of the Beam.io to a component I/O transparently if you decorated your PTransform with @PartitionMapper or @Processor.

The limitation are:

  • Inputs must implement PTransform<PBegin, PCollection<?>> and must be a BoundedSource.

  • Outputs must implement PTransform<PCollection<?>, PDone> and just register on the input PCollection a DoFn.

More information on that topic on How to wrap a Beam I/O page.

Advanced: define a custom API

It is possible to extend the Component API for custom front features.

What is important here is to keep in mind you should do it only if it targets not portable components (only used by the Studio or Beam).

In term of organization it is recommended to create a custom xxxx-component-api module with the new set of annotations.

Extending the UI

To extend the UI just add an annotation which can be put on @Option fields which is decorated with @Ui. All its members will be put in the metadata of the parameter. Example:

public @interface MyLayout {

Talend Component Testing Documentation

Best practises

this part is mainly around tools usable with JUnit. You can use most of these techniques with TestNG as well, check out the documentation if you need to use TestNG.

Parameterized tests

This is a great solution to repeat the same test multiple times. Overall idea is to define a test scenario (I test function F) and to make the input/output data dynamic.

JUnit 4

Here is an example. Let’s assume we have this test which validates the connection URI using ConnectionService:

public class MyConnectionURITest {
    public void checkMySQL() {
        assertTrue(new ConnectionService().isValid("jdbc:mysql://localhost:3306/mysql"));

    public void checkOracle() {
        assertTrue(new ConnectionService().isValid("jdbc:oracle:thin:@//myhost:1521/oracle"));

We clearly identify the test method is always the same except the value. It can therefore be rewritter using JUnit Parameterized runner like that:

@RunWith(Parameterized.class) (1)
public class MyConnectionURITest {

    @Parameterized.Parameters(name = "{0}") (2)
    public static Iterable<String> uris() { (3)
        return asList(

    @Parameterized.Parameter (4)
    public String uri;

    public void isValid() { (5)
1 Parameterized is the runner understanding @Parameters and how to use it. Note that you can generate random data here if desired.
2 by default the name of the executed test is the index of the data, here we customize it using the first parameter toString() value to have something more readable
3 the @Parameters method MUST be static and return an array or iterable of the data used by the tests
4 you can then inject the current data using @Parameter annotation, it can take a parameter if you use an array of array instead of an iterable of object in @Parameterized and you can select which item you want injected this way
5 the @Test method will be executed using the contextual data, in this sample we’ll get executed twice with the 2 specified urls
you don’t have to define a single @Test method, if you define multiple, each of them will be executed with all the data (ie if we add a test in previous example you will get 4 tests execution - 2 per data, ie 2x2)
JUnit 5

JUnit 5 reworked this feature to make it way easier to use. The full documentation is available at junit.org/junit5/docs/current/user-guide/#writing-tests-parameterized-tests.

The main difference is you can also define inline on the test method that it is a parameterized test and which are the values:

@ValueSource(strings = { "racecar", "radar", "able was I ere I saw elba" })
void mytest(String currentValue) {
    // do test

However you can still use the previous behavior using a method binding configuration:

void mytest(String currentValue) {
    // do test

static Stream<String> stringProvider() {
    return Stream.of("foo", "bar");

This last option allows you to inject any type of value - not only primitives - which is very common to define scenarii.

don’t forget to add junit-jupiter-params dependency to benefit from this feature.



component-runtime-junit is a small test library allowing you to validate simple logic based on Talend Component tooling.

To import it add to your project the following dependency:


This dependency also provide some mocked components that you can use with your own component to create tests.

The mocked components are provided under the family test :

  • emitter : a mock of an input component

  • collector : a mock of an output component

JUnit 4

Then you can define a standard JUnit test and use the SimpleComponentRule rule:

public class MyComponentTest {

    @Rule (1)
    public final SimpleComponentRule components = new SimpleComponentRule("org.talend.sdk.component.mycomponent.");

    public void produce() {
        Job.components() (2)
             .component("collector", "test://collector")

        final List<MyRecord> records = components.getCollectedData(MyRecord.class); (3)
        doAssertRecords(records); // depending your test
1 the rule will create a component manager and provide two mock components: an emitter and a collector. Don’t forget to set the root package of your component to enable it.
2 you define any chain you want to test, it generally uses the mock as source or collector
3 you validate your component behavior, for a source you can assert the right records were emitted in the mock collect
JUnit 5

The JUnit 5 integration is mainly the same as for JUnit 4 except it uses the new JUnit 5 extension mecanism.

The entry point is the @WithComponents annotation you put on your test class which takes the component package you want to test and you can use @Injected to inject in a test class field an instance of ComponentsHandler which exposes the same utilities than the JUnit 4 rule:

@WithComponents("org.talend.sdk.component.junit.component") (1)
public class ComponentExtensionTest {
    @Injected (2)
    private ComponentsHandler handler;

    public void manualMapper() {
        final Mapper mapper = handler.createMapper(Source.class, new Source.Config() {

                values = asList("a", "b");
        final Input input = mapper.create();
        assertEquals("a", input.next());
        assertEquals("b", input.next());
1 The annotation defines which components to register in the test context.
2 The field allows to get the handler to be able to orchestrate the tests.
if it is the first time you use JUnit 5, don’t forget the imports changed and you must use org.junit.jupiter.api.Test instead of org.junit.Test. Some IDE versions and surefire versions can also need you to install either a plugin or a specific configuration.
Mocking the output

Using the component "test"/"collector" as in previous sample stores all records emitted by the chain (typically your source) in memory, you can then access them using theSimpleComponentRule.getCollectoedRecord(type). Note that this method filters by type, if you don’t care of the type just use Object.class.

Mocking the input

The input mocking is symmetric to the output but here you provide the data you want to inject:

public class MyComponentTest {

    public final SimpleComponentRule components = new SimpleComponentRule("org.talend.sdk.component.mycomponent.");

    public void produce() {
        components.setInputData(asList(createData(), createData(), createData())); (1)

        Job.components() (2)
             .component("out", "yourcomponentfamily://myoutput?"+createComponentConfig())

1 using setInputData you prepare the execution(s) to have a fake input when using "test"/"emitter" component.
Creating runtime configuration from component configuration

The component configuration is a POJO (using @Option on fields) and the runtime configuration (ExecutionChainBuilder) uses a Map<String, String>. To make the conversion easier, the JUnit integration provides a SimpleFactory.configurationByExample utility to get this map instance from a configuration instance.


final MyComponentConfig componentConfig = new MyComponentConfig();
// .. other inits

final Map<String, String> configuration = configurationByExample(componentConfig);

The same factory provides a fluent DSL to create configuration calling configurationByExample without any parameter. The advantage is to be able to convert an object as a Map<String, String> as seen previously or as a query string to use it with the Job DSL:

final String uri = "family://component?" +

It handles the encoding of the URI to ensure it is correctly done.

Testing a Mapper

The SimpleComponentRule also allows to test a mapper unitarly, you can get an instance from a configuration and you can execute this instance to collect the output. Here is a snippet doing that:

public class MapperTest {

    public static final SimpleComponentRule COMPONENT_FACTORY = new SimpleComponentRule(

    public void mapper() {
        final Mapper mapper = COMPONENT_FACTORY.createMapper(MyMapper.class, new Source.Config() {{
            values = asList("a", "b");
        assertEquals(asList("a", "b"), COMPONENT_FACTORY.collectAsList(String.class, mapper));
Testing a Processor

As for the mapper a processor is testable unitary. The case is a bit more complex since you can have multiple inputs and outputs:

public class ProcessorTest {

    public static final SimpleComponentRule COMPONENT_FACTORY = new SimpleComponentRule(

    public void processor() {
        final Processor processor = COMPONENT_FACTORY.createProcessor(Transform.class, null);
        final SimpleComponentRule.Outputs outputs = COMPONENT_FACTORY.collect(processor,
                        new JoinInputFactory().withInput("__default__", asList(new Transform.Record("a"), new Transform.Record("bb")))
                                              .withInput("second", asList(new Transform.Record("1"), new Transform.Record("2")))
        assertEquals(2, outputs.size());
        assertEquals(asList(2, 3), outputs.get(Integer.class, "size"));
        assertEquals(asList("a1", "bb2"), outputs.get(String.class, "value"));

Here again the rule allows you to instantiate a Processor from your code and then to collect the output from the inputs you pass in. There are two convenient implementation of the input factory:

  1. MainInputFactory for processors using only the default input.

  2. JoinInputfactory for processors using multiple inputs have a method withInput(branch, data) The first arg is the branch name and the second arg is the data used by the branch.

you can also implement your own input representation if needed implementing org.talend.sdk.component.junit.ControllableInputFactory.


The folowing artifact will allow you to test against a spark cluster:

JUnit 4

The usage relies on a JUnit TestRule. It is recommended to use it as a @ClassRule to ensure a single instance of a spark cluster is built but you can also use it as a simple @Rule which means it will be created per method instead of per test class.

It takes as parameter the spark and scala version to use. It will then fork a master and N slaves. Finally it will give you submit* method allowing you to send jobs either from the test classpath or from a shade if you run it as an integration test.

Here is a sample:

public class SparkClusterRuleTest {

    public static final SparkClusterRule SPARK = new SparkClusterRule("2.10", "1.6.3", 1);

    public void classpathSubmit() throws IOException {
        SPARK.submitClasspath(SubmittableMain.class, getMainArgs());

        // do wait the test passed
this is working with @Parameterized so you can submit a bunch of jobs with different args and even combine it with beam TestPipeline if you make it transient!
JUnit 5

The integration with JUnit 5 of that spark cluster logic uses @WithSpark marker for the extension and let you, optionally, inject through @SparkInject, the BaseSpark<?> handler to access te spark cluster meta information - like its host/port.

Here is a basic test using it:

class SparkExtensionTest {

    private BaseSpark<?> spark;

    void classpathSubmit() throws IOException {
        final File out = new File(jarLocation(SparkClusterRuleTest.class).getParentFile(), "classpathSubmitJunit5.out");
        if (out.exists()) {
        spark.submitClasspath(SparkClusterRuleTest.SubmittableMain.class, spark.getSparkMaster(), out.getAbsolutePath());

        await().atMost(5, MINUTES).until(
                () -> out.exists() ? Files.readAllLines(out.toPath()).stream().collect(joining("\n")).trim() : null,
                equalTo("b -> 1\na -> 1"));
How to know the job is done

In current state, SparkClusterRule doesn’t allow to know a job execution is done - even if it exposes the webui url so you can poll it to check. The best at the moment is to ensure the output of your job exists and contains the right value.

awaitability or equivalent library can help you to write such logic.

Here are the coordinates of the artifact:


And here is how to wait a file exists and its content (for instance) is the expected one:

    .atMost(5, MINUTES)
        () -> out.exists() ? Files.readAllLines(out.toPath()).stream().collect(joining("\n")).trim() : null,
        equalTo("the expected content of the file"));


The HTTP JUnit module allows you to mock REST API very easily. Here are its coordinates:

this module uses Apache Johnzon and Netty, if you have any conflict (in particular with netty) you can add the classifier shaded to the dependency and the two dependencies are shaded avoiding the conflicts with your component.

It supports JUnit 4 and JUnit 5 as well but the overall concept is the exact same one: the extension/rule is able to serve precomputed responses saved in the classpath.

You can plug your own ResponseLocator to map a request to a response but the default implementation - which should be sufficient in most cases - will look in talend/testing/http/<class name>_<method name>.json. Note that you can also put it in talend/testing/http/<request path>.json.

JUnit 4

JUnit 4 setup is done through two rules: JUnit4HttpApi which is responsible to start the server and JUnit4HttpApiPerMethodConfigurator which is responsible to configure the server per test and also handle the capture mode (see later).

if you don’t use the JUnit4HttpApiPerMethodConfigurator, the capture feature will be deactivated and the per test mocking will not be available.

Most of the test will look like:

public class MyRESTApiTest {
    public static final JUnit4HttpApi API = new JUnit4HttpApi();

    public final JUnit4HttpApiPerMethodConfigurator configurator = new JUnit4HttpApiPerMethodConfigurator(API);

    public void direct() throws Exception {
        // ... do your requests

For tests using SSL based services, you will need to use activeSsl() on the JUnit4HttpApi rule.

If you need to access the server ssl socket factory you can do it from the HttpApiHandler (the rule):

public static final JUnit4HttpApi API = new JUnit4HttpApi().activeSsl();

public void test() throws Exception {
    final HttpsURLConnection connection = getHttpsConnection();
    // ....
JUnit 5

JUnit 5 uses a JUnit 5 extension based on the HttpApi annotation you can put on your test class. You can inject the test handler (which has some utilities for advanced cases) through @HttpApiInject:

class JUnit5HttpApiTest {
    private HttpApiHandler<?> handler;

    void getProxy() throws Exception {
        // .... do your requests
the injection is optional and the @HttpApi allows you to configure several behaviors of the test.

For tests using SSL based services, you will need to use @HttpApi(useSsl = true).

You can access the client SSL socket factory through the api handler:

@HttpApi*(useSsl = true)*
class MyHttpsApiTest {
    private HttpApiHandler<?> handler;

    void test() throws Exception {
        final HttpsURLConnection connection = getHttpsConnection();
        // ....
Capturing mode

The strength of this implementation is to run a small proxy server and auto configure the JVM: http[s].proxyHost, http[s].proxyPort, HttpsURLConnection#defaultSSLSocketFactory and SSLContext#default are auto configured to work out of the box with the proxy.

It allows you to keep in your tests the native and real URLs. For instance this test is perfectlt valid:

public class GoogleTest {
    public static final JUnit4HttpApi API = new JUnit4HttpApi();

    public final JUnit4HttpApiPerMethodConfigurator configurator = new JUnit4HttpApiPerMethodConfigurator(API);

    public void google() throws Exception {
        assertEquals(HttpURLConnection.HTTP_OK, get("https://google.fr?q=Talend"));

    private int get(final String uri) throws Exception {
        // do the GET request, skipped for brievity

If you execute this test, it will fail with a HTTP 400 because the proxy doesn’t find the mocked response. You can create it manually as seen in the introduction of the module but you can also set the property talend.junit.http.capture to the folder where to store the captures. It must be the root folder and not the folder where the json are (ie not prefixed by talend/testing/http by default).

Generally you will want to use src/test/resources. If new File("src/test/resources") resolves to the valid folder when executing your test (Maven default), then you can just set the system property to true, otherwise you need to adjust accordingly the system property value.

Once you ran the tests with this system property, the testing framework will have created the correct mock response files and you can remove the system property. The test will still pass, using google.com…​even if you disconnect your machine from the internet.

The rule (extension) is doing all the work for you :).

Passthrough mode

Setting talend.junit.http.passthrough system property to true, the server will just be a proxy and will execute each request to the actual server - like in capturing mode.

Beam testing

If you want to ensure your component works in Beam the minimum to do is to try with the direct runner (if you don’t want to use spark).

Check beam.apache.org/contribute/testing/ out for more details.

Multiple environments for the same tests

JUnit (4 or 5) already provides some ways to parameterized tests and execute the same "test logic" against several data. However it is not that convenient to test multiple environments.

For instance, with Beam, you can desire to test against multiple runners your code and it requires to solve conflicts between runner dependencies, setup the correct classloaders etc…​It is a lot of work!

To simplify such cases, the framework provides you a multi-environment support for your tests.

It is in the junit module and is usable with JUnit 4 and JUnit 5.

JUnit 4

public class TheTest {
    public void test1() {
        // ...

The MultiEnvironmentsRunner will execute the test(s) for each defined environments. It means it will run test1 for Env1 and Env2 in previous example.

By default JUnit4 runner will be used to execute the tests in one environment but you can use @DelegateRunWith to use another runner.

JUnit 5

JUnit 5 configuration is close to JUnit 4 one:

class TheTest {

    void test1() {
        // ...

The main difference is you don’t use a runner (it doesn’t exist in JUnit 5) and you replace @Test by @EnvironmentalTest.

the main difference with JUnit 4 integration is that the tests are execute one after each other for all environments instead of running all tests in each environments sequentially. It means, for instance, that @BeforeAll and @AfterAll are executed once for all runners.

Provided environments

The provided environment setup the contextual classloader to load the related runner of Apache Beam.

Package: org.talend.sdk.component.junit.environment.builtin.beam

the configuration is read from system properties, environment variables, …​.
Class Name Description



Contextual runner



Direct runner



Flink runner



Spark runner

Configuring environments

If the environment extends BaseEnvironmentProvider and therefore defines an environment name - which is the case of the default ones, you can use EnvironmentConfiguration to customize the system properties used for that environment:

    environment = "Direct",
    systemProperties = @EnvironmentConfiguration.Property(key = "beamTestPipelineOptions", value = "..."))

    environment = "Spark",
    systemProperties = @EnvironmentConfiguration.Property(key = "beamTestPipelineOptions", value = "..."))

    environment = "Flink",
    systemProperties = @EnvironmentConfiguration.Property(key = "beamTestPipelineOptions", value = "..."))
class MyBeamTest {

    void execute() {
        // run some pipeline
if you set the system property <environment name>.skip=true then the environment related executions will be skipped.
Advanced usage
this usage assumes Beam 2.4.0 is in used and the classloader fix about the PipelineOptions is merged.



These dependencies brings into the test scope the JUnit testing toolkit, the Beam integration and the multi-environment testing toolkit for JUnit.

Then using the fluent DSL to define jobs - which assumes your job is linear and each step sends a single value (no multi-input/multi-output), you can write this kind of test:

class TheComponentTest {
    void testWithStandaloneAndBeamEnvironments() {
        // add asserts on the output if needed

It will execute the chain twice:

  1. with a standalone environment to simulate the studio

  2. with a beam (direct runner) environment to ensure the portability of your job

Secrets/Passwords and Maven

If you desire you can reuse your Maven settings.xml servers - including the encrypted ones. org.talend.sdk.component.maven.MavenDecrypter will give you the ability to find a server username/password from a server identifier:

final MavenDecrypter decrypter = new MavenDecrypter();
final Server decrypted = decrypter.find("my-test-server");
// decrypted.getUsername();
// decrypted.getPassword();

It is very useful to not store secrets and test on real systems on a continuous integration platform.

even if you don’t use maven on the platform you can generate the settings.xml and settings-security.xml files to use that feature. See maven.apache.org/guides/mini/guide-encryption.html for more details.

Generating data?

Several data generator exists if you want to populate objects with a semantic a bit more evolved than a plain random string like commons-lang3:

A bit more advanced, these ones allow to bind directly generic data on a model - but data quality is not always there:

Note there are two main kind of implementations:

  • the one using a pattern and random generated data

  • a set of precomputed data extrapolated to create new values

Check against your use case to know which one is the best.

an interesting alternative to data generation is to import real data and use Talend Studio to sanitize the data (remove sensitive information replacing them by generated data or anonymized data) and just inject that file into the system.

If you are using JUnit 5, you can have a look to glytching.github.io/junit-extensions/randomBeans which is pretty good on that topic.

Scroll to top