CustomJavaComponent
Short Description |
Ports |
Metadata |
CustomJavaComponent Attributes |
Details |
Public CloverDX API |
Examples |
Best Practices |
Compatibility |
See also |
Short Description
CustomJavaComponent executes user-defined Java code.
Component | Same input metadata | Sorted inputs | Inputs | Outputs | Each to all outputs | Java | CTL | Auto-propagated metadata |
---|---|---|---|---|---|---|---|---|
CustomJavaComponent | - | - | 0-n | 0-n | - |
Ports
Number of ports depends on the Java code.
Metadata
By default, CustomJavaComponent does not propagate metadata. Metadata propagation can be configured using the Metadata propagation algorithm and Metadata propagation algorithm class attributes (see also the Custom Metadata Propagation example).
CustomJavaComponent has no metadata templates.
Requirements on metadata depend on user-defined transformation.
CustomJavaComponent Attributes
Attribute | Req | Description | Possible values |
---|---|---|---|
Basic | |||
Algorithm | [1] | A runnable transformation in Java defined in a graph. | |
Algorithm URL | [1] | A URL to an external file defining a runnable transformation in Java. | |
Algorithm class | [1] | An external runnable transformation class. | |
Advanced | |||
Metadata propagation algorithm | An algorithm for metadata propagation written in Java. | ||
Metadata propagation algorithm class | An external metadata propagation class. | ||
Algorithm source charset | The encoding of the external file defining the transformation. The default encoding depends on DEFAULT_SOURCE_CODE_CHARSET in defaultProperties. | UTF-8 | |
[1] One of these must be set. These transformation attributes must be specified. |
Details
CustomJavaComponent executes Java transformation.
There are specialized custom java components: CustomJavaReader, CustomJavaWriter and CustomJavaTransformer. These components differ just in the provided Java template.
You can use Public CloverDX API in this component. See Public CloverDX API.
External JAR Files
The default folder for external .jar
files in a local project is ./lib
.
On the Server, external .jar
files can also be placed on the classpath of the Worker.
You should add the .jar
files to classpath
.
Open Project Properties dialog
→ .
Switch to
→ .
Click and select the .jar
files.
Running on Cluster
All .java
and .class
files should reside in a shared sandbox.
Editing Code in Another Tab
If you click the Algorithm
or Metadata propagation algorithm attribute value,
a dialog for editing of build-in java code opens.
Use the Switch to Java editor button
to convert the transformation in Java to a .java
file.
The file is opened as a new tab having Java editor with syntax highlighting associated.
Java Interfaces for CustomJavaComponent
Transformation required by the component must extend the
org.jetel.component.AbstractGenericTransform
class.
Following are methods of the AbstractGenericTransform
class:
ConfigurationStatus checkConfig(ConfigurationStatus status)
Use this method to check the configuration of a custom component: custom attributes and their values, ports and metadata.
void execute()
Define your transformation here. The method is called once when the component is started.
void init()
Initializes Java class/function. This method is called only once at the beginning of the transformation process. Any object allocation/initialization should happen here.
void preExecute()
This is also an initialization method, which is invoked before each separate graph run. Contrary to the
init()
procedure, only resources for this graph run should be allocated here. All resources allocated here should be released in thepostExecute()
method.void postExecute()
This is a de-initialization method for a single graph run. All resources allocated in the
preExecute()
method should be released here. It is guaranteed that this method is invoked after a graph finish, at the latest. For some graph elements, for instance components, this method is called immediately after a phase finish.Node getComponent()
Returns an instance of component. Useful for accessing component attributes.
File getFile(String fileUrl)
Returns a file for a given file URL.
InputStream getInputStream(String fileUrl)
Returns
InputStream
for a given file URL.OutputStream getOutputStream(String fileUrl, boolean append)
Returns
OutputStream
for a given file URL.
Custom Metadata Propagation for CustomJavaComponent
CustomJavaComponent can propagate metadata using a custom algorithm
(see the example).
Metadata propagation algorithm must extend the
org.jetel.component.GenericMetadataProvider
class.
Following are methods of the GenericMetadataProvider
class:
void propagateMetadata()
Define your metadata propagation algorithm here. Use the
setInputMetadata()
andsetOutputMetadata()
methods to propagate metadata to connected ports. ThepropagateMetadata()
method may be called multiple times before the graph is executed. Results of any expensive operations should be cached (e.g. in instance fields). Any resources opened in this method must also be closed in this method.DataRecordMetadata getMetadataFromInputPort(int portIndex)
Returns an instance of metadata that is currently propagated from outside to input port.
DataRecordMetadata getMetadataFromOutputPort(int portIndex)
Returns an instance of metadata that is currently propagated from outside to output port.
void setInputMetadata(int portIndex, DataRecordMetadata metadata)
Assigns given metadata to input port of this component.
void setOutputMetadata(int portIndex, DataRecordMetadata metadata)
Assigns given metadata to output port of this component.
Node getComponent()
Returns an instance of component. Useful for accessing component attributes.
File getFile(String fileUrl)
Returns a file for a given file URL.
Public CloverDX API
Data Record |
Data Field |
Metadata |
Dictionaries |
Lookup Tables |
Graph Parameters |
Component Attributes |
Sequences |
Database Connections |
Opening Streams |
Logging |
Public CloverDX API is a set of CloverDX Java classes you can use in transformations in CustomJavaComponent and other components using Java transformation.
Public CloverDX API uses the @CloverPublicAPI
annotation.
Classes annotated by @CloverPublicAPI
are part of the API
and can be used in your transformation.
Details on particular classes are documented in javadoc.
The following pieces of code serve to point to particular classes suitable for a particular purpose.
You can use the standard Java classes and classes provided by the API in your transformations. Do not use CloverDX Java classes not included in the API! The classes not included in the API may be changed in the next release, or removed.
Data Record
One single record is represented by the DataRecord
class.
DataRecord record;
To create a data record not connected to any particular port,
use the static method newRecord()
of DataRecordFactory
.
It requires record metadata.
String metadataId = getGraph().getDataRecordMetadataByName("recordName1");
DataRecordMetadata metadata = getGraph().getDataRecordMetadata(metadataId);
DataRecord record = DataRecordFactory.newRecord(metadata);
To read a record from the input port, use readRecordFromPort()
function.
The index of the input port (starting from 0) is specified in the parameter of the function.
record = readRecordFromPort(0);
The function returns null
if no other records are available.
while ((record = readRecordFromPort(0)) != null) { // Do something }
To write a record to the output port, use the writeRecordToPort()
function.
Parameters define the index of the output port and record to be written.
writeRecordToPort(0, record);
If you create a component with a variable number of the input or output ports, use
getComponent().getInputPortsMaxIndex()
or
getComponent().getOutputPortsMaxIndex()
to get the maximal index of input or output ports.
Note that the first input port has index 0. A component with N input ports has N-1 as the maximal index of input port.
Data Field
Data field is represented by the DataField
class.
You can get fields of data record using getFields()
of DataRecord
.
You can get a particular field using getField()
taking the field index or field name as a parameter.
DataRecord record;
DataField dataField1;
dataField1 = record.getField(0);
DataField dataField2;
dataField2 = record.getField("field2");
Use getValue()
and setValue()
methods
of DataField
to work with field values.
DataField field = record.getField(0);
String value = field.getValue();
field.setValue("some new value");
Metadata
There are two classes necessary to work with metadata:
DataRecordMetadata
and DataFieldMetadata
.
DataRecordMetadata
represents metadata of the whole record
whereas DataFieldMetadata
represents metadata of particular field.
Use getMetadata()
method of DataRecord
to get access to metadata of a record.
DataRecord record; record = ... DataRecordMetadata metadata = record.getMetadata();
Use getMetadata()
method of DataField
to get access to metadata of a field.
DataField dataField; dataField = ... DataFieldMetadata fieldMetadata = dataField.getMetadata();
To use metadata depending on its name, use getDataRecordMetadataByName()
to get metadata id and subsequently use getDataRecordMetadata()
to get metadata corresponding to the id.
String metadataId = getGraph().getDataRecordMetadataByName("recordName1");
DataRecordMetadata metadata = getGraph().getDataRecordMetadata(metadataId);
Dictionaries
To read a value from a dictionary, use the getValue()
function,
to write to a dictionary use the setValue()
function:
Dictionary dictionary = getGraph().getDictionary(); dictionary.setValue("mykey1", "NewValue"); String s = dictionary.getValue("mykey2");
Lookup Tables
The LookupTable
interface gives you an access to lookup tables.
Use put()
to insert a data record into an existing lookup table.
Note that getLookupTable()
requires lookup table ID.
The parameter is not the lookup table name!
LookupTable lookup = getGraph().getLookupTable("LookupTable0");
DataRecord record = ...;
lookup.put(record);
Use createLookup()
to search for items matching the key.
LookupTable lt; lt = ... DataRecord patternRecord = DataRecordFactory.newRecord(lt.getMetadata()); patternRecord.getField(0).setValue("keyToBeSearchedPart1"); patternRecord.getField(2).setValue("keyToBeSearchedPart2"); String [] lookupFields = {"field1", "field3"}; RecordKey recordKey = new RecordKey(lookupFields, lt.getMetadata()); Lookup lookup; lookup = lt.createLookup(recordKey, patternRecord); lookup.seek(); while (lookup.hasNext()){ DataRecord record = lookup.next(); // process the result found writeRecordToPort(0, record); }
Graph Parameters
Graph parameters can be obtained from TransformationGraph
using getGraphParameters()
.
To get a particular parameter use getGraphParameter()
with the parameter name.
TransformationGraph graph = getGraph();
GraphParameters graphParameters = graph.getGraphParameters();
GraphParameter graphParameter = graphParameters.getGraphParameter("MY_PARAMETER");
Use getValue()
or getValueResolved()
to get the parameter value.
String value = graphParameter.getValue(); String valueResolved = graphParameter.getValueResolved(RefResFlag.REGULAR);
Component Attributes
You can get a value of component attributes
using the getProperty()
functions applied on TypedProperties
.
String myStringValue = getProperties().getStringProperty("myCustomPropertyName1");
Integer myIntegerValue = getProperties().getIntProperty("myCustomPropertyName2");
Sequences
A sequence is accessible from TransformationGraph
via the getSequence()
function with the sequence ID as a parameter.
Sequence seq = getGraph().getSequence("Sequence0");
Use nextValueInt()
, nextValueLong()
or nextValueString()
to increment the sequence and return the incremented value.
A first call of any of the nextValue*()
functions
initializes the sequence to the initial sequence value and returns an unincremented initial value.
String sequenceValue = seq.nextValueString(); int sequenceValueInt = seq.nextValueInt(); long sequenceValueLong = seq.nextValueLong();
To get the last value returned by functions above use currentValueInt()
currentValueLong()
or currentValueString()
.
If none of the nextValue*()
functions have been called before,
the current value is the start value of the sequence.
String sequenceValue = seq.currentValueString(); int sequenceValueInt = seq.currentValueInt(); long sequenceValueLong = seq.currentValueLong();
Database Connections
Database connections are accessible via the getDBConnection()
method
of AbstractGenericTransform
.
The method requires connection name or connection ID as a parameter.
Connection connection = getDBConnection("myUniqueID");
The getDBConnection(String)
method is available since 4.7.0-M2.
The access to database connection in earlier versions was different.
JNDI
To connect to JNDI data source from Custom Java Component, create a database connection using the JNDI data source. Use this connection in your source code in the same way as in the case of connecting to a database without a JNDI data source. See the example above.
Opening Streams
If you work with paths, use the getFile()
function to resolve the path correctly.
String param = getProperties().getStringProperty("InputFile");
File file = getFile(param);
You can access files via streams.
Use getOutputStream()
or getInputStream();
String param = getProperties().getStringProperty("InputFile");
InputStream is = getInputStream(param);
String param = getProperties().getStringProperty("OutputFile"); OutputStream os = getOutputStream(param, true);
Logging
Use the log()
function to log messages of important events
of your Java-defined transformation.
getLogger().log(Level.INFO, "Some message" );
Examples
Remover of Empty Directories |
Checking Configuration of Custom Component |
Custom Metadata Propagation |
Remover of Empty Directories
Create a component removing empty directories.
Solution
Add a new attribute Directory to the component.
Use the following code.
package jk; import java.io.File; import org.jetel.component.AbstractGenericTransform; /** * This is an example custom component. It shows how you can remove empty * directories. */ public class CustomJavaComponentExample01 extends AbstractGenericTransform { private void removeEmptyDirectories(File dir) { if (!dir.isDirectory() || !dir.canRead() || !dir.canWrite()) { return; } for (File f : dir.listFiles()) { if (f.isDirectory()) { removeEmptyDirectories(f); if (f.listFiles().length == 0) { f.delete(); } } } } @Override public void execute() { String directory = getProperties().getStringProperty("Directory"); File dir = getFile(directory); removeEmptyDirectories(dir); } }
Checking Configuration of Custom Component
The component has to have one input port and one output port connected. Each port should have metadata assigned. The component has the attribute Multiplier having integral value.
Solution
Use the checkConfig()
function of a component's template.
@Override public ConfigurationStatus checkConfig(ConfigurationStatus status) { super.checkConfig(status); if (getComponent().getInPorts().size() != 1 || getComponent().getOutPorts().size() != 1) { status.add("One input and one output port must be connected!", Severity.ERROR, getComponent(), Priority.NORMAL); return status; } DataRecordMetadata inMetadata = getComponent().getInputPort(0).getMetadata(); DataRecordMetadata outMetadata = getComponent().getOutputPort(0).getMetadata(); if (inMetadata == null || outMetadata == null) { status.add("Metadata on input or output port not specified!", Severity.ERROR, getComponent(), Priority.NORMAL); } if (!getProperties().containsKey("Multiplier")) { status.add("Multiplier property is missing or is not set.", Severity.ERROR, getComponent(), Priority.NORMAL, "Multiplier"); return status; } try { Integer.parseInt(getProperties().getStringProperty("Multiplier")); } catch(Exception e){ status.add("Multiplier is not integer!", Severity.ERROR, getComponent(), Priority.NORMAL, "Multiplier"); } return status; }
Custom Metadata Propagation
Below is an example of the code
which adds an additional Date
field to any metadata coming from input,
and propagates the enhanced metadata to output port:
DataRecordMetadata metadataFromInputPort = getMetadataFromInputPort(0); // Look up metadata from input port 0. if (metadataFromInputPort != null) { // Always check for null, the metadata may not be available initially. metadataFromInputPort.setName(metadataFromInputPort.getName() + "_modified"); // Modify the name of the metadata. // Add additional field to the metadata that came from input port. DataFieldMetadata additionalField = new DataFieldMetadata("CreatedDate", DataFieldType.DATE, null); additionalField.setFormatStr("yyyy-MM-dd"); // Configure the format property of our new field. metadataFromInputPort.addField(additionalField); setOutputMetadata(0, metadataFromInputPort); // Propagate the modified metadata to output port. }
Best Practices
If the transformation is specified in an external file (with Algorithm URL), we recommend users to explicitly specify Algorithm source charset.
Compatibility
Version | Compatibility Notice |
---|---|
4.1.0-M1 | CustomJavaComponent is available since 4.1.0-M1. It replaced JavaExecute. |
5.3.0 | CustomJavaComponent can now propagate metadata - added new attributes Metadata propagation algorithm and Metadata propagation algorithm class. |