Handling "No Data" Entries in PI System with RDBMS Interface

Handling "No Data" Entries in PI System with RDBMS Interface

When interfacing with an RDBMS to pull data into your PI System, encountering "No Data" entries can be a puzzling yet crucial part of data management. Understanding why this happens and how to manage it can significantly enhance the integrity and clarity of your data analysis.

Why "No Data" Appears

When using an RDBMS interface, the presence of "No Data" in PI System typically traces back to NULL values in your source SQL database—in this case, a WinCC database. Before actual values are finalized, the table might contain placeholder NULL values. This can lead to temporary "No Data" entries in PI as the interface correctly reflects the current state of database records.

In practice, this behavior prevents outdated or unfinalized test results from being mistakenly interpreted as current data. Particularly in environments where timely data updates are critical (like industrial or operational settings), showing "No Data" until full results are available ensures a clear, accurate representation of what is known at each data point in time.

Managing the "No Data" Entries

Instead of disabling "No Input Errors," which would suppress this informative state, consider configuring the interface to replace "No Data" with actual values once they become available. This is largely managed through the Location5 parameter in your interface configuration:

• Without /rbo or /lb flags: Set Location5 to -1 to enable the replacement of "No Data" entries with actual values when they share the same timestamp. This setting instructs the PI System to overwrite the "No Data" state with concrete data as soon as it appears in the database.

• With /rbo flag: Current configurations do not support replacement, maintaining "No Data" for strictly bound-order reads.

• With /lb flag: This configuration allows for replacement regardless of Location5 value, facilitating flexible data overwrite capabilities at matching timestamps.

Configuring for Better Outcomes

To implement these changes effectively:

1. Understand Your Data Needs: Before setting any configurations, be clear whether your data workflows need to display interim "No Data" states or if they should always replace old data with new as soon as possible.

2. Set the Correct Location5 Parameter: Review the RDBMS interface configuration manual and adjust the Location5 parameter based on your use of /rbo or /lb.

3. Monitor and Verify: After configuration, monitor your PI trends to ensure data behaves as expected. Spot-check timestamps to verify that "No Data" states are correctly replaced.

By understanding and manipulating how your RDBMS interface handles NULL values and timestamps, you can greatly enhance the accuracy and utility of your PI System data.

Conclusion

Handling "No Data" scenarios in the PI System should be seen as an opportunity to fine-tune data flows and ensure system accuracy. By adjusting the Location5 settings thoughtfully, users can better align their data presentation with practical operational needs.