PDREP

PDREP post-processes a report produced by PowerDesigner to reduce the noise level by removing pieces with no real information, and compacts the information about keys and index to a single list for each. You need at least version 9.5 of PowerDesigner to use PDREP.

Running PDREP

 pdrep report

When you generate a report from PowerDesigner you get a file, say MyReport.html, and a directory MyReport_files which includes a number of files called MyRe1.html, MyRepo2.html etc plus some system files for stylesheets, navigation in the menu tree etc. When you run PDREP, PDREP creates new files in this directory. Each file covers a single table or a single data type etc and is named after the object covered in the file. The only file generated by PowerDesigner that is replaced is the table of contents (TOC). The old TOC is renamed by adding .old as an extension, for instance Report_toc.html.old. If you run PDREP twice on the same report file, PDREP will detect this, and do nothing the second time.

If there is a file footer.html in the directory when you run PDREP, PDREP will add this footer to all files it generates. This can be useful for copyright notes.

When you run PDREP, it prints the name of the file it is currently producing.

If PDREP fails to parse an HTML file produced by PowerDesigner, PDREP stops. In this case, PDREP may have generated one or more files, but the original TOC is unmolested, as the TOC files are not renamed unless PDREP completes successfully.

Report Templates

To use PDREP, you must use any of the two report templates that comes with AbaPerls and which you find in the Misc directory of your AbaPerls installation:

The two templates works largely the same, but htmlrep16.rtp has a couple of enhancements:

• Support for included columns in indexes.
• Any text in the Notes tab for the index in included to the index list, so that you can document why a certain index exists.
• If you add options on the Microsoft tab for the index properties, they are displayed in a column. This is also used for filtered indexes, see below.
• The template htmlrep16.rtp addresses a change in PowerDesigner 16 how the Clustered property for keys is handled.
• htmlrep951.rtp only includes sections for user-defined types and tables, while htmlrep16.rtp also includes sections for views and stored procedures.
• For foreign-key columns there are columns to show any action for the key (CASCADE, SET NULL or SET DEFAULT).

If you wish to extend the reports you get from the template, you are free to try. But don't be surprised, if PDREP crashes or botches the extra report items you added. And you must in no way change the order of the columns in the tables in the template. Also be careful not to change the order in some sections. For instance, PDREP assumes that the individual section for indexes are in the same order as in the Index List.

Report Settings

It is also important to have the correct settings in PowerDesigner when you generate the report. Select Report->Report properties from the menu in the report-edit window. Here is a screen short from PowerDesigner 16.1:

There are several changes you need to from the defaults. The most important are highlighted in the screen.shot. If you do not change these, PDREP will not work.

Here is a complete list of recommendations:

• On the tab Title Page select that you don't want any title page; this page is of little use. But if it's there, PDREP will cope with it.
• Template and Stylesheet. Use the template and style sheet that appeals to you. Any template should work with PDREP.
• Header. There is little reason to include any of the header that comes with PowerDesigner, but you may want to add your own logo here or some such.
• Footer. Use this option if you think the footer is useful. You can also add a footer with a footer.html when run PDREP, see above.
• Home page. I advice that you leave this empty.
• Table of contents. Leave this at the default.
• Image format. This does not affect PDREP.
• List format. You must set these to high numbers, or else PowerDesigner may split the description of a table with many columns over several HTML files, which PDREP does not cope with.
• Web browser. I have only tested with HTML5, but I would expect the other setting to work as well.
• Add page break for top level lists and cards. You must uncheck this option. The templates include page breaks where this is desired. PDREP will crash on the first file, if you don't uncheck the option.

Not all these options may be available in older verisons of PowerDesigner, why you don't have to worry about them with one exception. In earlier versions of PowerDesigner, the List format settings were not exposed in the GUI, but you must set them in the registry according to this table:

Workaround for Filtered Indexes

While PowerDesigner 16.1 claims to support SQL 2008, it does not have support for filtered indexes. To this end, both PDREP and TBLFIX support a workaround. On the proprerty page for an index there is a tab Microsoft. If you put the index filter in the Filegroup box, as shown in the figure below, both PDREP and TBLFIX will interpret this as an index filter.

Note! You must include WHERE in the text, or else both tools will assume that you did indeed mean a file group.

Modifications performed by PDREP

Here follows a brief summary of the most important modifications performed by PDREP:

• The TOC is modified so that for the table descriptions there is one node per letter. Tables with names ending in names are placed in a node at the end.
• In the column list, the primary-key columns are in bold face.
• The identity property is displayed in the NOT NULL column as IDNT.
• The column list for keys and indexes are included in the Keys and Indexes table. Included columns are added here as well, as are index options.
• The clustered index (which may be a key) is cast in bold face.
• Columns with no interesting data is removed from the Keys and Indexes tables as from the tabels for foreign keys.
• The report generated by PowerDesigner seem to always to include a Server Check Expression for all columns, even though in many cases it reads %MINMAX% and %LISTVAL% and %CASE% and %RULES%. There are also always a table with Standard checks. PDREP removes all such entries with no real information.
• PDREP removes the ### entered for table constraints. (This is a special convention used by TBLFIX to permit you to have multiple table-level constraints for a table.

To see an example of the output from PDREP, see the documentation for the AbaPerls system tables. This report was generated from the files Abaperls.pdm and footer.html in the doc/database directory of your AbaPerls installation.