When you get JSON strings and you need to extract a few fields to table format for Spotfire consumption it would be nice having an efficient and easy to use parser to do this out of the box.
Something like:
JSONGetField( JSONString, ”afieldname” )
where “afieldname” should allow addressing nested JSON elements such as
“address.zipcode”
or
“orderlines[3].productname”
and, why not, even aggregations and statistical calculations such as
sum(JSONGetFields( JSONString,”orderlines[*].price” ) )
where the "*" could be even further elaborated to support filters (ranges, first, last, and other filters).
The parser should decide internally if extracting just the field on demand or, if the user is extracting quite a few fields from the same JSON string, pre-parse the full string to a machine efficient format and then efficiently address the fields that the user is extracting. This should help with performance when loading thousands of JSON records with many field extractions from them.