Generate Alerts Using a Derived Window in SAP HANA Streaming Analytics

Intermediate

20 min.

SAP HANA streaming analytics, Intermediate, Internet of Things, SAP HANA, express edition, Tutorial

Create and use a derived window to analyze events, filter data, and generate alerts.

You will learn

How to use a derived window to apply a filter to watch for events that should generate an alert

Prerequisites

Tutorials: Create a Moving Average on an Event Stream using an Aggregation Window

In this tutorial, you will add and configure a derived stream that will watch for events that need to generate an alert - by applying a filter.

However, rather than just using the standard Filter element, you will use the Derived Window element from the Palette since it gives you more flexibility. With this, you can apply the filter and also change the structure of the output event.

At this point in the tutorial series, if you don’t want to go through the process of creating this window using the visual editor, then you can simply switch to the CCL editor ( F6 ) and paste the CCL below into your project. From there, you can then skip to the last step to compile and test your project.


CREATE OUTPUT WINDOW ALARM_TEMP
PRIMARY KEY DEDUCED
KEEP ALL
AS SELECT
  AVG_TEMP.MACHINEID MACHINEID ,
  AVG_TEMP.EVENT_TIME EVENT_TIME ,
  AVG_TEMP.AVG_TEMP AVG_TEMP ,
  AVG_TEMP.MAX_TEMP MAX_TEMP ,
  AVG_TEMP.LOCATION LOCATION ,
  'TEMPERATURE' ALARM_TYPE ,
  'Machine not maintaining' ALARM_DESC
FROM AVG_TEMP
WHERE AVG_TEMP.AVG_TEMP > AVG_TEMP.MAX_TEMP ;

Step 1
Click on Derived Window in the Palette and drop it onto the canvas.
Step 2
Add a connector from AVG_TEMP to the new window.
With this new window added to your freezer_monitor project, based on the work you’ve done so far in this tutorial and the previous tutorial group (Get started with SAP HANA streaming analytics for SAP HANA, express edition), how many elements (not including connector arrows) are now on your canvas?
7
6
8
9
Step 3
Change the new window name to ALARM_TEMP, then press Enter.
Step 4
Click the Add Column Expression ( f(x) ) option shown below.
Choose Copy Columns from Input from the menu to open the column selection dialog.
Select all except for MIN_TEMP and TEMP_UNIT, then click OK.
Select the Add Column Expression ( f(x) ) option, then Column Expression to add a column to this window.
Repeat this step to add a second column.
Step 5
Rename the first new column to ALARM_TYPE, then press Enter.
Rename the second new column to ALARM_DESC.
Double-click on simpleResultExpression6 to edit the expression.
Enter 'TEMPERATURE' in the expression edit box for the ALARM_TYPE column. This will set the “type” of all alarms emitted by this window to the string “TEMPERATURE”. Press Enter to confirm your entry.
Note that it’s important to enclose string values in single quotes in expressions.
Enter 'Machine not maintaining temperature' in the expression edit box for the ALARM_DESC column. Press Enter to confirm your entry.
Step 6
Now click the Add Query Clause button.
Choose Filter from the menu or press f.
You want to add a filter so that this window only contains rows for machines that have a current average temperature above the max specified for the machine.
Double-click on 1 under Other to edit the filter expression.
Change the filter expression to: AVG_TEMP.AVG_TEMP > AVG_TEMP.MAX_TEMP. Use Ctrl+Space for completion assist.
Step 7
Click the Compile Streaming Project (F7) button to check for errors.
Note: You will get a compiler warning about the lack of a retention policy on the aggregate stream. You can ignore this warning. The reason for the warning is that windows default to KEEP ALL. If you explicitly state KEEP ALL, then the compiler accepts that you meant it. If you omit an explicit KEEP policy, the compiler warns you in case it’s not what you intended. Here, this is an aggregation where you are grouping by Machine ID - and you do indeed want to keep the aggregate value for every machine ID.