Being primarily focused on using the application (rather than programming it), tracing can be a great benefit to helping explain why something does not work as you expected or to help with understanding how the application retrieved particular data you see. Below are some common scenarios where tracing can be quite helpful:
- See where prompt fields are getting their data.
- Find out the records used to store a transaction (employee, vendor, account, code, etc.).
- Using tracing to help create queries or reports.
- Understanding Batch Applications with trace files.
- Other Uses.
How many times have you been using a page and then clicked that magnifying glass icon and were surprised by the result? Maybe you feel you have configured some values but they are not showing up, or there are values showing up you did not expect. The screenshot below provides an example of four such prompt fields:
Although there are many ways to find out exactly how PeopleSoft is retrieving the list, the simplest is usually to merely do a SQL trace of clicking on the icon to find out what PeopleSoft is actually requesting from the database. Follow the steps below to get to this information:
- Login to PeopleSoft and navigate to the page with the prompt field.
- Click the New Window link to open a new web browser page.
- Turn on online tracing using the Interactive Tracing technique.
- Back on the original page, click the magnifying glass by the field you are researching.
- Turn off tracing (again, see Interactive Tracing technique).
- Load the trace file into Pace-Trace. Although you can leave all the checkboxes clicked in Pace-Trace, only SELECT statements will be of interest.
Since only the statements involved in bringing up the list (even if no items were returned) will be shown in the trace file, you now have the critical information you need to continue your research.
For example, suppose a new account you added was not displayed when you clicked to list available accounts on a different page. After tracing, you found the following SQL:
SELECT ACCOUNT, DESCRAfter looking back to the page where accounts are defined, you notice that your account was not defined as an expense account, and make the correction. Rarely is it this simple, but use this approach as a model.
WHERE STATUS = A
AND TYPE = E
By turning on tracing just before you click the Save icon for the first time for any particular online entry (be it configuration, like an Account, Job Code, or Vendor; or a transaction, such as a Voucher or Purchase Order), you can use the trace file contents to list all records that are written to the database. This information can be very helpful later to build online queries or for creating specification documents.
To trace an online save, turn on tracing just before clicking the Save icon or simply trace an entire session; when you open the trace file in Pace-Trace, unclick the checkboxes by SELECT and leave checked the UPDATE, DELETE, and INSERT checkboxes.
For many reports and queries, the easiest way to find out what records to use will be to use the process above to find records used to store an online entry; however, this approach has a few problems with it:
- It does not produce output in a convenient SELECT format that is used in the SQL tab of PS Query (which you can use to verify your query).
- Some pages that consolidate all the information you require are view only, and therefore offer no means to trace a save of this data.
- Although you should always have a test database that allows you to save test (or junk) records, this may not always be desired (or possible).
By tracing the SQL that is produced by opening a page, you can find out exactly where this information is stored. Follow the instructions for turning on tracing to turn on tracing just before opening the page online.
Be prepared for tons of SQL, however, as there are sometimes hundreds of SQL statements involved in displaying a page online. To minimize the excess SQL in your file, make sure you have opened this page recently without tracing turned on- there is a caching process the application server must do the first time it opens a page which adds many SQL statements to the file (in addition to slowing your entire process). After that verification, open the page again with tracing turned on and review the resulting trace file.
By reviewing SQL trace files created from the execution of batch processes, you can determine what records are being created and what changes are made to your existing transactions to mark them as complete. You can take this information and use it to learn how to identify processed transactions, and understand the destination of this processing. After turning on tracing, try to process as few (but representative) transactions as possible, keeping the trace file small, and your research less taxing.
Tracing can be used to help with learning new functionality, creating specification documents for developers, documenting how something changed (performing tracing in two different environments), and simply troubleshooting all kinds of problems. Once you are quick with turning on and off tracing for both online and batch activity, you will continually find uses for it.