How to read a SQL Trace in PeopleSoft


As always with so many things, I struggled initially to understand the trace file. More often, I would generate a trace file and did not understand every aspect of the trace file. However with experience, I have figured what is written to the trace file and sharing my findings below.

When a trace is run for SQL statements, the resulting trace statement will have various parts. Here is the list along with the description.

First Part: n-xxxxx.
This is a sequential line counter for the process (exe). n is a integer starting from 1 to n. xxxxx is reserved for each line written to the trace file. If there is a second process, then it would be 2-xxxxx, for the third one it would be 3-xxxxx and so on.

Second Part: hh.mi.ss
It indicates the timestamp at which the trace line is written. This timestamp is retrieved from the machine in which PeopleTools is running.

Third Part: A time value
This is the time elapsed since the previous trace line was written. The time elapsed between n-xxxxx and n-(xxxxx+1) is written.

Fourth Part: Cur#n
This indicates the cursor number for the statement

Fifth Part: PSFT_DB
Indicates the PeopleSoft database in which this API call is executing.

Sixth Part: RC=0
This is the return code for the associated API call.

Seventh Part: Dur=Another time value.
This is the time to execute the assoicated API call.

Eight Part: COM Stmt=<SQL Statement>
This is the database API call and provides information on the SQL executed.

Next post is on reading a Peoplecode trace.

3 comments:

  1. Trailing in Peoplecode often result in crumming of the SQL statement. Integer multiparted through the SQL... Visit : www.mypeoplesoft.com for further information.

    ReplyDelete
  2. Now we have one tool which make easy to read the trace file .You can find the tool in Oracle metalink site .Tools name is "TraceMagic".

    ReplyDelete

Please refrain for marketing messages and unnecessary back links.