Introduction

DataVision is an Open Source reporting tool similar to Crystal Reports. Reports can be designed using a drag-and-drop GUI. They may be run, viewed, and printed from the application or exported as HTML, XML, PDF, LaTeX2e, DocBook, or tab- or comma-delimited text files. The output files produced by LaTeX2e and DocBook can in turn be used to produce PDF, text, HTML, PostScript, and more. DataVision is written in Java and runs almost anywhere. It can generate reports from databases or text data files. Any database with an available JDBC driver should work: Oracle, PostgreSQL, MySQL, Informix, hsqldb, Microsoft Access, Progress, and more. Columns read from text files can be separated by any character. Report descriptions are stored as XML files. This means you can not only use the DataVision GUI but you may also edit reports using your favorite text editor. DataVision is developed and maintained by Jim Menard (jimm@io.com). The latest version of DataVision can be found on the DataVision Web page. New releases are also announced on Freshmeat and on the DataVision mailing list.

Recent Changes

Here is the summary of the major changes in version . For a complete list of changes, see the ChangeLog file. Changes

DataVision now requires Java 1.4.X. (I don't know if it will work with Java 1.5.)

New aggregate fields, replacing the single ``subtotal'' aggregate field. Field type values may be ``sum'', ``min'', ``max'', ``count'', ``average'', and ``stddev''. The old ``subtotal'' type will still be accepted and will be turned into ``sum''. The old ``Running Total'' dialog now lets you select the aggregate type at each level (group or report footer).

The file parsing API has been simplified. Almost everywhere in the public interface, such as the report reading methods, that used to take a String file path, a Reader, an InputStreamReader, an InputSource, or anything else now just takes one of two types: a File or a InputSource.

The old readFile() methods, that were deprecated, have been removed. Use the read() equivalents. See the previous item, though: read() only takes a File or an InputSource.

The old Report.setParameterXMLFile method, that was deprecated, has been removed. Use setParameterXMLInput, which now only takes a File or an InputSource.

Stopped using the MinML2 SAX parser, in favor of the built-in Java SAX parser.

Removed the GNU regex library, since Java 1.4 includes regex support.

Upgraded to iText PDF generation library version 1.3.

Upgraded to JRuby 0.8.2.

Added Rick Kramer's patch which adds a delete menu item to the field picker window's edit menu.

Stop using ``enum'' as a variable name. It's a reserved word as of Java 1.5.

The class ConnectionTest in examples/ConnectionTest.java is now in the package named ``examples''. The class TranslateOMatic in jimm/properties/TranslateOMatic.java is now in the package ``jimm.properties''.

I've switched to using Subversion (SVN) instead of CVS for source code control. SourceForge will be offering SVN in the future. For now, I'm hosting the repository at http://jimm.textdriven.com/svn/datavision/trunk. I haven't bothered to try to import the existing CVS repository, since I don't think anybody really cares about the complete revision history of all the code since I've checked it in to CVS. To check out the source anonymously, assuming you have a SVN client, type
```
svn co http://jimm.textdriven.com/svn/datavision/trunk datavision
```
which will create a new directory called datavision. You can also point your browser at that URL and browse the latest revision online.
Remember the last-used layout engine in the export window. This is the first of a hopefully long line of things that DataVision will remember for you. I plan to add more, like the windows positions, last-opened report, default formats, paper sizes, database connections, and more.

Bug Fixes

Fixed creation of datavision-versiontar.gz. When unarchived, the files appear in a subdirectory named datavision-version instead of being unpacked in the current directory.
Fixed a bug in the GUI: when a field format was changed, then the report was run, then another field's format was changed, the changed format was not used when you ran the report again. (The format was saved correctly when you saved the report.)
Fixed bug where saving a report didn't save it to the right directory; it was saved in the datavision dir.
Improved the PDF layout engine's font selection. It wasn't using the correct font name when it looked for the PDF equivalent standard font. The Java font family name is used to pick the PDF font. The family can be the font name ("Helvetica", "Arial") or a logical name ("sansserif", "serif", "monospace"). "Helvetica" or "sansserif" map to Helvetica, "Courier" or "monspace" map to Courier, "Symbol" and "ZapfDingbats" map to themselves, and everything else (including "serif", for example) maps to Times.