Plateau Report Designer was created with the Report Author in mind. Plateau selected BIRT as the underlying engine for its power and flexibility. As a result, the skill set required to author reports grows as the complexity of the report grows.
But no matter how skilled you are, you should expect to spend some time troubleshooting the report. We have built in troubleshooting help, including logging (set from elms.xml) and the ability to run unpublished reports (you can run the report from Talent Management before the report goes live. During troubleshooting, keep the following in mind:
Common Plateau Report Designer Problems: Common issues that arise and solutions to them.
Troubleshooting Report Designer SQL: The SQL query is the heart of report writing. This section contains information specific to the SQL statement in the reports.
BIRT Behavior to Avoid: Because Plateau Report Designer is built as a plug-in to BIRT, it retains some BIRT behavior that is less useful, or in some cases can cause problems during maintenance stages of the report.
The following are common issues with Plateau Reports and how you can solve them.
By default, the BIRT plug-in for Eclipse sets a small number of tables to return in the data set. You can change the preference at Windows > Preferences > Report Design > Data Set Editor > JDBC Data Set > Maximum number of tables in each schema to display. We recommend setting it to 100. Avoid setting prefetch to true, especially if you are working in a development database.
If you cannot see any libraries, such as parameters, export the libraries from Talent Management. Talent Management acts as the system of record for libraries (and labels). For more information about what you need when you first start Plateau Report Designer, see Getting Started.
Queries can break on import for many different reasons, so check the logs as described in Troubleshooting Report SQL. However, common problems with SQL queries include:
BIRT Joins: BIRT joins (done in the Plateau Report Designer interface rather than in the SQL itself) can cause performance problems because the BIRT engine makes the join in memory. We recommend authoring your joins in the SQL itself because the DBMS is better suited for handling joins.
Double-Clicking in Data Set: If you double click an entity in the data set, Plateau Report Designer adds the SID info to the Query. We recommend typing the SQL.
Test Parameters: If you create test parameters with dummy data to test your report, and you forget to remove the dummy data and parameters when you import, you might have trouble importing the report or if the report imports, it might not run properly. Remove all dummy data and test parameters before you import.
The ApplicationResources.properties file in your Plateau Report Designer workspace must match the ApplicationResources.properties file in the BIRT resources folder (defined in elms.xml).
Coordinate with the Talent Management administrator to match the Talent Management .properties file with the .properties file in your workspace. You can provide the administrator with your .properties file and request that he or she import the file using the import/export labels wizard. Doing so prompts Talent Management to update the BIRT resources folder (defined in elms.xml) for you.
You can import your reports in the unpublished state. Unpublished reports are accessible only to a select number of users: generally users who are administering and testing reports. While reports are in the unpublished state, you can communicate with the administrator to properly import, and test, your reports.
To debug BIRT report queries, set the Log4J reporting level to DEBUG and search the logged SQL output from the reports. (BIRT engine log settings are stored in the elms.xml file). Talent Management writes the SQL to the log (or console) and you can see the run-time SQL.
Talent Management adds comments to the SQL to help you find what you are looking for. There are three different comments added, and each has a start and an end comment.
SEARCH FILTER: The filtering search SQL
SEC FILTER: The domain filtering (security) SQL
ORDER BY: The ordering statement SQL
For example, at run time, you see:
select a.stud_id, a.lname, a.fname
from pa_student a
where a.active = some selection
/* SEARCH FILTER START*/
and a.stud_id in (filtering query)
/* SEARCH FILTER END*/
/* SEC FILTER START*/
and 1=1
/* SEC FILTER END*/
/*ORDER BY START*/
order by order by field
/*ORDER BY END*/
Sometimes, you see 1=1 in one of the filters. Talent Management substitutes 1=1 for a SQL statement when a condition is always true.
Because Plateau Report Designer is built as a plug-in to BIRT, you can do things in BIRT that are not wrong, but are contrary to the spirit and purpose of Plateau Report Designer.
In Plateau Report Designer, you can filter report data in one of two ways:
You can use BIRT functionality to build filters. Your SQL query returns numerous records to memory and the BIRT filter removes unwanted records before the results appear to the user.
You can use a WHERE clause in your SQL statement to filter records. Your SQL query returns only the records that you want, resulting in a lesser hit to memory.
For performance reasons, write SQL statements that only return records that you want in your report. It is much more efficient to return only records that will appear in your report. Doing so puts the processing cost on the host most suited for the task of sorting through data: the Database.
In general, you always want to write a better SQL statement than use BIRT functionality that replicates something you could write in a SQL query.
When you drag the Data Sets from the Data Explorer and drop it to the report workspace, the data binding works between the table and the data set. This might not be your intention. For example, during maintenance, you might change the Data Set or the table in ways that upset the intention of the report.
Instead of dragging the Data Set and dropping it to the workspace, consider inserting the columns in the table one by one. Doing so binds the column to the data element itself.
In general, watch out for the way your data is binding.
Sometimes, BIRT inserts the SID into your report. For example, in the Edit Data Set > Query area BIRT, if you double-click a column in the Available Items area, BIRT adds it to the query, but it does so with the SID.
When the SID is in the query itself, it can cause problems when you migrate from development to staging to production.
In general, be aware of anything in your query that is schema-specific.
Avoid spaces in parameter names because in some instances, the spaces create problems with Talent Management functionality. For example, spaces in date parameters cause the date picker icon to disappear.