Amend RML test cases to improve parity with R2RML counterparts

[mcndt]

Over in the r2rml-test-cases-support repo, some R2RML test cases have been amended because they resulted in failing tests depending on how the test is implemented.

While I noticed that in most cases the RMLTC equivalents of these tests already had their expected outputs changed to avoid these test failures, now that this is fixed in the R2RMLTC cases perhaps the RMLTC equivalents should be amended such that they are the same test case as their R2RML counterpart.

Involved cases:

• via Fix inconsistent pairings of mapping files and expected output: 0005b, 0012a, 0012e r2rml-test-cases-support#17
  • RMLTC0005b
  • RMLTC0012a
  • RMLTC0012e
• via Change order of table drops for running tests sequentially with a MySQL source r2rml-test-cases-support#16
  • RMLTC0009{x} (a quick look at the resources in this repo shows that this issue was already amended in the RMLTC versions of the test case, so afaik nothing needs to be changed for this one)