Ungridded signal export for oData workaround

[Shamus Cunningham]

Currently in Seeq (June 2021) there is not a method to export ungridded signal data. There are a couple of development items in the works which will help in the future but for the moment this is the best work around for oData connections. As a reminder if you just need the data in excel, there is an ungridded raw timestamp option available for the excel export. 

Lets use the following setup as our example. We have a condition which represents my batches and we have a simple temperature signal that I want to get the average value of for each batch. 

Step 1 - Create my Average temperature per batch KPI using the Signal from Condition tool. In this example we are selecting the start of the batch to place the KPI timestamp. 

Step 2 - Re-grid my discrete points onto a known fixed grid. This step will loose some fidelity with the batch start timestamps and is the main drawback of this method. Pick a re-grid interval that ensures that you will not have two points inside the same interval (hours, minutes, days, weeks) 

$AveCycleTemp.aggregate(average(), hours(), startkey())

Step 3 - Step export on the same new grid by selecting Custom grid period and Days, minutes, hours to match your re-grid interval from Step 2. Select the OData Sample Table endpoint when you create the export. 

Step 4 - One final optional step to clean things up in your BI tool is to remove all the null points when you import the oData feed. The example below shows the process in the PowerBI Power query editor but there should be similar steps in other tools.