Data tracking considerations

Before you implement data tracking in a process application, make sure you understand the supported data types and naming conventions, as well as any considerations for working with versioned data.

Supported data types

Data types that IBM® BPM tracks include the following:

Type of tracking Supported data types
Autotracking String, Integer, Decimal, Boolean, and Date
Tracking groups String, Number, and Date

When you are tracking data, be aware of the following:

  • A tracking group can have a maximum of 50 fields for each of the data types. Do not increase the maximum number of fields because it can affect the database table size and future migration to a new IBM BPM release.
  • Variables for which the Is List option is enabled cannot be tracked.
  • Complex types cannot be mapped directly; their fields must be mapped individually.
  • Variables of type ANY, Map, Record, XMLDocument, XMLElement, and XMLNodeList cannot be tracked.

Naming tracking groups

When naming tracking groups and tracked fields, be aware of the following restrictions:

  • Do not use SQL-92 reserved words. Several sources available on the internet provide a complete list of SQL-92 reserved words.
  • Do not use any of the names used for the views and fields in the Business Performance Data Warehouse database schema.

Tracking data across processes and process applications

To track data from multiple processes (BPDs) that reside in the same process application, create a tracking group and implement it for as many BPDs as you like, mapping the tracked fields to the appropriate variables for each BPD.

If you want to capture data from multiple processes (BPDs) that reside in different process applications, you can do so by using the same tracking group in each process application. For example, you can create a tracking group in a toolkit, and then create a dependency on that toolkit in each process application where you want to use the tracking group. From each process application, you can implement the tracking group one or more times, mapping the tracked fields to variables within each application. When you send tracking definitions and then run instances of the BPDs, the data is captured in a single tracking group view as described in "Business Performance Data Warehouse tracking group views." The data that IBM BPM captures enables you to analyze the tracked data in any way you choose. For example, you can analyze the tracked fields as a whole or you can compare the data from each process application or from each process.

Working with versioned data

All data tracked by IBM BPM includes snapshot (version) information that enables you to create reports to compare versions of your processes if you have that requirement.

When tracking data, keep the following in mind:

  • Timing intervals work across snapshots (versions). For example, a process that starts in one version (1.0) might be migrated to a new version (2.0) before reaching the end of a timing interval. In such a case, the data for the timing interval is captured in the Performance Data Warehouse as expected with the version change noted.
  • Data types of variables that are being stored (autotracked or part of a tracking group) can change between versions. If data types do change, a new column is created in the corresponding view as described in "Performance Data Warehouse database architecture."