CA Mainframe Application Tuner (CA MAT) collects SQL data using two different methods.
One is based on observing the target address space and the other is based on the harvester.
The harvester is the more exact method of the two because it monitors every SQL statement that gets processed in the DB2 subsystem.
The observation and sampling method is the less accurate of the two. But both are done depending on what is specified for DB2 options.
After the data is collected for both options and put it into the monitor file during analysis the data will be merged based on things like DBRM name, CRC value, statement number, etc., and DB2view will then show the merged data.
If on the DB2View screen only SQL from the sample process is, there will be an S in the 'Data From' column.
If it is only from the Harvester there will be an H in the 'Data From' column and if there is SQL in both sample and harvester there will be a B in the 'Data From' column.