For example in 2i, the source is defined with rml:query "SELECT Name FROM student";, selecting only column Name. Then in a term map, the column ID is also used: rr:template "http://example.com/{ID}/{Name}";, which throws an error.
This is not the goal of the respective test cases however, which test for missing SQL version identifiers and the use of qualified column names.
(I think this slipped under the radar because missing SQL version identifiers were not properly implemented in the mapper, see here and here.)
I propose to simplify the mappings in these test cases so they only test what is in their descriptions (SQL version identifiers/qualified column names), but use column names which areSELECTed. SELECTing column names is already partially covered by test cases 2g and 2h.
For example in 2i, the source is defined with
rml:query "SELECT Name FROM student";
, selecting only columnName
. Then in a term map, the columnID
is also used:rr:template "http://example.com/{ID}/{Name}";
, which throws an error.This is not the goal of the respective test cases however, which test for missing SQL version identifiers and the use of qualified column names.
(I think this slipped under the radar because missing SQL version identifiers were not properly implemented in the mapper, see here and here.)
I propose to simplify the mappings in these test cases so they only test what is in their descriptions (SQL version identifiers/qualified column names), but use column names which are
SELECT
ed.SELECT
ing column names is already partially covered by test cases 2g and 2h.This leads to an error in the test case report generator, see https://github.com/kg-construct/rml-test-cases-support/issues/9