All Toolkit scripts MUST return columns in a consistent order between subsequent samples. If they do not then bad things can happen. |
---|
|
We have seen external scripts, especially when working with un-ordered data structures like that returned as JSON, will do a straight forward conversion to the Toolkit's CSV format without any regard for the ordering of the columns, instead relying on the source data to conform. This has a negative effect on the Toolkit plugin which expects the columns to be in a consistent order. If the columns change, whether through add or deletes or changes to the order, then the plugin has to do a complex "dance" under the covers to reconfigure the Dataview and the changes being sent upstream to the Gateway. Another consequence of this is that some data cells will be momentarily incorrectly changed (this is not normally visible to a human user) that in turn may trigger Rules unexpectedly and the data values being logged or published will have the wrong values while the Dataview re-converges to the correct values. Data Charts will also show unexpected peaks or troughs when the data value changes are large. This will also apply to the order of Headline values. |
|
Comments
0 comments
Please sign in to leave a comment.