Please find below some questions in regards to Power BI Mapping.
Number of records in Power BI sometimes doesn't match the one in the triplestore.
By your experience, what could be causing such a problem? Sometimes, just by refreshing the data in Power BI causes the numbers to change when the source is a static dataset.
I tried to add reasoning to the Power BI connection by following the instructions here, item 4, and got the following error: Expression.Error: The field 'Data' of the record wasn't found.
Any recommendation on how to map bnodes? Do you have any in-depth documentation on BI Mapping?
I read your question before you edited and I could see the issue. You need to include the "optional" property in the field mappings as documented in SQL Schema Field Mapping Options. Non-optional fields will constrain the underlying SPARQL query (they are not NULLable and required to be present).
To put it another way, if you map a class "Person" and add a field "name", it is not optional by default. Any query on the Person table will return only rows in which the SPARQL query has a value for name, whether it is requested in the SQL query or not. By setting the field to be optional, the data type is now NULLable and it's not required for the corresponding SPARQL query to return only Person entities which have name triples.
I hope this helps your first issue. Please share some details if you're still blocked.
Hi Jess, yes, the optional worked for most things. We can say it resolved the first issue, thanks so much. I do see the number of records changing on the Power BI side when i refresh the data. It shouldn't as the data in the triplestore doesn't change between refreshes. I need to keep an eye on it and documented it properly when it happens again.