Question

I have a simple Java program that takes a .jrxml file, compiles it, then fills it. The data for the report is supplied in an XML file. After the fill is done, the data is exported to PDF.

// Parse input document
Document document = JRXmlUtils.parse(new File(xmlFile));

// Set it as the data source in the parameters
parameters.put(JRXPathQueryExecuterFactory.PARAMETER_XML_DATA_DOCUMENT, document);

// Create and set the virtualizer
JRFileVirtualizer virtualizer = new JRFileVirtualizer(2, "/tmp");
virtualizer.setReadOnly(true);
parameters.put(JRParameter.REPORT_VIRTUALIZER, virtualizer);

// Fill the report
String jasperFile = designFile.replaceAll(".jrxml",".jasper");
print = JasperFillManager.fillReport(jasperFile, parameters);

// Export the report to PDF
ArrayList<JasperPrint> jasperPrints = new ArrayList<JasperPrint>();
jasperPrints.add(print);
JRPdfExporter exp = new JRPdfExporter();
exp.setParameter (JRExporterParameter.JASPER_PRINT_LIST, jasperPrints);
exp.setParameter(JRExporterParameter.OUTPUT_FILE_NAME, outFileName );
exp.exportReport();

My confusion is about the export

That fill line above works great for small reports, but once I get an XML source file approaching 1/2 MB, it will spin for over a day, regardless of setting the file virtualizer (which I do).

I see there is another method called fillReportToSteam. My confusion is that with fillReport, I have to do an extra step to export to PDF. What sort of stream is fillReportToStream writing to, and how do I specify that? Will fillReportToStream write to a PDF file?

I haven’t been able to find any examples. I was hoping I could take advantage of the stream so that I could measure the progress and get these PDFs to complete in a normal span of time.

Answer 1

Calling fillReportToStream still produces a JasperPrint object, but it writes it directly to an output stream instead of passing it back to you. The JasperFillManager cannot output the report in any other format, so the export step cannot be skipped. To generate a PDF, you would still need to use JRPdfExporter and read the filled report from wherever the output stream wrote it to.

I think the problem here is your file virtualizer. You mention that “it will spin for over a day, regardless of setting the file virtualizer” but the virtualizer will actually make the process take longer. It is a basic time/memory tradeoff to avoid out of memory errors, but will make the filling much slower. In a benchmark I’ve seen, adding a file virtualizer quadrupled the filling time!

If you really need the virtualiser, try increasing the maxSize parameter you are passing. 2 seems awfully low. You could also try switching to a JRSwapFileVirtualizer as I’ve heard they have much better performance.