.:Documentation and tutorials
 Project page
Vote for iReport on O'Reilly's OSDir.com

<< Previous   |   Next >>   |   Table of Contents
7 Using the XML datasource

Starting from version 0.3.2, iReport provides an XML datasource that permit to fill a report reading data from an XML file.
iReport 0.4.0 use JRXmlDatasource shipped with JasperReports (from version 0.6.x)

This new implementation was developed by Peter Severin (developer of JasperAssistant).
This new implementation is much powerful supporting real XPath expressions. Here is what JasperReports documentation says about JRXmlDatasource.

Class JRXmlDataSource

All Implemented Interfaces:
JRDataSource, JRRewindableDataSource

public class JRXmlDataSource
extends java.lang.Object
implements JRRewindableDataSource
XML data source implementation that allows to access the data from a xml document using XPath expressions.

The data source is constructed around a node set (record set) selected by an XPath expression from the xml document.

Each field can provide an additional XPath expresion that will be used to select its value. This expression must be specified using the "fieldDescription" element of the field. The expression is evaluated in the context of the current node thus the expression should be relative to the current node.

To support subreports, sub data sources can be created. There are two different methods for creating sub data sources. The first one allows to create a sub data source rooted at the current node. The current node can be seen as a new document around which the sub data source is created. The second method allows to create a sub data source that is rooted at the same document that is used by the data source but uses a different XPath select expression.


 	<B id="0">
 	<B id="1">
 	<D id="3">

Data source creation

  • new JRXmlDataSource(document, "/A/B") - creates a data source with two nodes of type /A/B
  • new JRXmlDataSource(document, "/A/D") - creates a data source with two nodes of type /A/D
Field selection
  • @id - will select the "id" attribute from the current node
  • C - will select the value of the first node of type C under the current node.
Sub data source creation
  • "((net.sf.jasperreports.engine.data.JRXmlDataSource)$P{REPORT_DATA_SOURCE}).subDataSource("/B/C") - in the context of the node B, creates a data source with elements of type /B/C
  • "((net.sf.jasperreports.engine.data.JRXmlDataSource)$P{REPORT_DATA_SOURCE}).dataSource("/A/D") - creates a data source with elements of type /A/D

Generally the full power of XPath expression is available. As an example, "/A/B[@id > 0"] will select all the nodes of type /A/B having the id greater than 0. For a short XPath tutorial check the following URL:

Peter Severin (peter_p_s@sourceforge.net, contact@jasperassistant.com)

This chapter cover the old iReport implementation of XML datasource. This can help to use the new one.

7.1 Configuring an XML datasource

The first step to use an XML file as a datasource is configure an XML Datasource. Select menu Datasource > Connections / Datasources, press New and select "XML file datasource" as connection type.
Fig.7.1: The XML datasource configuration panel.

Fill the name field with the name of this datasource (i.e. Adressbook), select the XML file and set the record path.

7.2 The Record Path

An XML document is structured as a tree. JasperReports need data organized as a record set, so we must present a tree as set of record. The Record Path is a string that define the minimum path required to define a record.
Example 1:
It this very simple case, the Record Path is the path from the root (tag palette) to the leaf (tag color): /palette/color
We use the same record path for this document (Example 2):
Please note that in this case the path don't ends with a leaf tag.
In general, the end tag is the first tag that has children of different type or is a leaf.

When the document become more complex, we can read prodoce a sub-datasource starting from a child. In example 3 we can see that any color can have 0 or more aliases...
	<alias type="html">#FF0000</alias>
	<alias type="rgb">255, 0, 0</alias>
	<alias type="CMYK">0, 99, 100, 0</alias>
	<alias type="html">#0000FF</alias>
	<alias type="rgb">0, 0, 255</alias>
	<alias type="lab">88, -79, 81</alias>
	<alias type="CMYK">63, 0, 100, 0</alias>
	<alias type="html">#0000FF</alias>
	<alias type="rgb">0, 0, 255</alias>
In this case we can think to tags alias as a set of records childs of master record /palette/color, the childs record will have as record path the value /color/alias:
    <alias type="html">#FF0000</alias>
    <alias type="rgb">255, 0, 0</alias>
    <alias type="CMYK">0, 99, 100, 0</alias>
Note as this XML is very similar to example 1.

7.3 XML Datasource Syntax

To bind fields inside the XML, a syntax similar to XPath is used. The field path is stored in the field description (this because the field name don't support a name like "/palette/color").
The path syntax is really simple. Is a recursive expression:

<base_path>[<symbol> <base_path>[<symbol><...>] ]

<base_path> is the path from the root to the tag. If this path is longer than the Record Path, it's cuted to the Record Path. It is required for a field definition. If it's alone, the value of the tag is returned.
Example: /palette/color
<symbol> is used to add an extra path to the base path and to define what should be returned.
+ add the following path to the base_path (this happen when the base_path = record path);
@ return the attribute value: it's followed by the attribute name;
* return all tags identified by the following path as a JRXMLDatasource



7.4 Addressbook sample

This sample shows how to use the JRXMLDatasource using subreports.
The sample is composed by 4 files:

addressbook.xml is the data file
addressbook.jrxml is the master report file
hobby.jrxml is the hobby subreport
eamil.jrxml is the email subreport


   <category name="home">
      <person id="1">                                                           
          	<hobby>Radio Control</hobby>
          	<hobby>R/C Cars</hobby>
          	<hobby>Micro R/C Cars</hobby>
          	<hobby>Die-Cast Models</hobby>
      <person id="2">
      <person id="3">                                                                 
          	<hobby>Science Hobby</hobby>
          	<hobby>Toy Horse</hobby> 
  <category name="work">    
      <person id="4">                                                            
          	<hobby>Toy Horse</hobby> 
      <person id="5">
      <person id="6">
      <person id="7">
  <category name="Other">    
      <person id="8">   
      <person id="9">

We have defined a datasource named Addressbook that use addressbook.xml as data file. The Record path is: /addressbook/category/person

Fields in the report are defined as follow:

 Field Name   Field Description   Type 
 CATEGORY   /addressbook/category@name   java.lang.String 
 PERSON_ID   /addressbook/category/person@id   java.lang.String 
 LASTNAME   /addressbook/category/person+LASTNAME   java.lang.String 
 FIRSTNAME   /addressbook/category/person+FIRSTNAME   java.lang.String 
 HOBBIES   /addressbook/category/person+hobbies*hobby   java.lang.Object (castable to JRDataSource) 
 EMAIL_ADDRESSES   /addressbook/category/person*email   java.lang.Object (castable to JRDataSource) 

To the master report (figure 7.2) we have added a group named CATEGORY that has as expression the category name.
Fig.7.2: The master report.
We have two subreport too, one for hobbies of a person and one for email addresses.
The first subreport element has $F{HOBBIES} as connection expression (of type datasource expression) and as subreport expression "/hobby.jasper" of type String.
The second subreport element has $F{EMAIL_ADDRESSES} as connection expression (of type datasource expression) and as subreport expression "/email.jasper" of type String.
In hobby.jasper we have defined only one field:
 Field Name   Field Description   Type 
 HOBBY   /hobbies/hobby   java.lang.String 

The same in email.jasper:
 Field Name   Field Description   Type 
 EMAIL   /person/email   java.lang.String 

Note as every field in subreport are defined specifing as root element in the path the parent element:
HOBBY -> /hobbies/hobby
EMAIL -> /person/email

This is the result:
Fig.7.2: The final result.
Finally we show how to use the JRXMLDataDource in a program:

1   package it.businesslogic.ireport.examples;
3   import it.businesslogic.ireport.connection.JRXMLDataSource;
4   import dori.jasper.engine.export.JRPdfExporter;
6   import dori.jasper.engine.*;
7   import java.util.HashMap;
10  public class XMLDataSourceExample {
12      public static void main(String[] args) throws Exception {
14                  String reportFileName = "/addressbook.jasper";
15                  String outFileName = "/addressbook.pdf";
16                  String xmlFileName = "/addressbook.xml";
17                  String recordPath = "/addressbook/category/person";
19                  JRXMLDataSource jrxmlds = new JRXMLDataSource(xmlFileName,recordPath);
21          HashMap hm = new HashMap();
23          try
24          {
25              JasperPrint print = JasperFillManager.fillReport(
26                          reportFileName, 
27                          hm, 
28                          jrxmlds);
30              JRExporter exporter = new dori.jasper.engine.export.JRPdfExporter();
32              exporter.setParameter(JRExporterParameter.OUTPUT_FILE_NAME,outFileName);
33                      exporter.setParameter(JRExporterParameter.JASPER_PRINT,print);
35                      exporter.exportReport();
36                      System.out.println("Created file: " + outFileName);             
37          }
38          catch (JRException e)
39          {
40              e.printStackTrace();
41              System.exit(1);
42          }
43          catch (Exception e)
44          {
45              e.printStackTrace();
46              System.exit(1);
47          }
49      }
51  }

7.5 XML Datasource license

JRXMLDatasource is not released under GPL license (as iReport is) but under LGPL as jasperreports: in this way you can use the JRXMLDataSource in your program without problems.
However, if you use it, please make a simbolic donation to the iReport project.

(c) 2002 by Giulio Toffoli <gt78 at businesslogic.it>