When working with PowerGREP, you will save your data in a number of different files. File selections are saved in *.pgf files, action definitions are saved in *.pga files, libraries are saved in *.pgl files, results are saved in *.pgr files and the undo history is saved in a *.pgu file. All these files are XML files. You can easily open them in a text editor or XML editor to look at their contents.
The main benefit of the XML format is that you can use standard XML software to read files saved by PowerGREP, or even create your own. The ability to create file selection files and action definition files is particularly useful. While PowerGREP offers a wide range of command line parameters, not all aspects of the file selection and action definition can be controlled from the command line. The flat command line is simply too cumbersome to control PowerGREP's wide range of abilities. A structured XML file is much more useful. Instead of controlling individual settings via command line parameters, simply use your favorite XML software or XML programming library to generate a .pgf and/or .pga file, and pass those on the command line to PowerGREP.
Reading PowerGREP results files can be very handy if you want to apply some special processing to the results found by PowerGREP. The XML structure of a *.pgr file stores all the information that PowerGREP itself uses to display the results on the Results panel, without requiring access to the files that were searched through to produce the results.
All five file formats used by PowerGREP are based on a single XML Schema. You can download the schema definition file from http://www.powergrep.com/powergrep50.xsd. When creating file selection and action files by yourself, make sure to validate them against the schema. PowerGREP does not validate files against the schema, and makes little effort to display helpful error messages.
In order to avoid inconsistencies, there is no separate documentation for PowerGREP's file formats. The XML schema is annotated, and serves as both human-readable documentation and machine-readable specification.
The XML schema is laid out in a bottom-up fashion. The root element, which is always "powergrep", is defined as the last element in the schema. If you collapse all nodes under xsd:schema you will get an overview of all the types the schema defines.