I have previously used a design pattern for querying some linked DBs which creates a carefully curated prepared statement and then runs the statement using the following steps:
Above, the commented out statement is what would be returned from the geoCodeStatement data.
In previous implementations, this EXEC query would return the table data and any columns I would more specifically define.
Today, however, when implementing this pattern, the returned data was not the table data but just a generic execution response:
Older apps which use this pattern are not returning the same response, but the actual table data (example below):
Was there recently an internal change for how executed statements are processed?