I have exported device data using 2 mechanisms:
- Data Explorer -> Export
- Execute notebook -> Download device_data.csv
Using Data Explorer, I get a column named “Device ID”. I noticed that a Losant Notebook I’ve written was throwing key errors, failing to detect a column named “Device ID”; downloading device_data.csv from the failed Notebook execution reveals that device_data.csv contains a column called “ID”.
So it seems that the device-ID column is named “Device ID” when Data Explorer is used, but the column is named “ID” when device_data.csv is delivered to the Notebook.
I think this is a bug, although I could be misunderstanding.
Good catch @Alexander_Farley!
The Data Explorer export format was something we hadn’t considered when choosing the format for Notebooks. We made it consistent with the existing device data export.
For anyone else running into this issue, we recommend using the notebook’s data export functionality to guarantee the data you’re working with locally is identical to what the notebook will receive at run time.
We definitely can’t change the format for notebooks since it will impact existing data pipelines, but we might be able to change the data explorer’s format to be consistent. We’ll have to do some thinking internally to understand the implications of making that change and how it’ll affect people depending on that format.
Thanks for finding this issue and bringing it to our attention!