Datasets, themselves, allow just about anything as a column type. But that freedom creates huge problems for Ignition's infrastructure when arbitrary datatypes are stored in tags. So the tag machinery of Ignition checks the columns types of any incoming dataset against a whitelist, and rejects any dataset that doesn't conform.
You have a couple options:
-
convert the troublesome columns to some other datatype (stringification is common, if you don't need the functionality of the object's true type), or
-
store the dataset in memory, typically in a top-level script library variable, and have your uses look for it there. (You might find my
globalVarMap()functions in my Integration Toolkit module helpful.)