René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Bad results for failing standalone expectations on compound data ","articleBody":"**Describe the bug**\r\nWhen running standalone expectations in a single session and having more than one failing expectation, the results returned by report are wrong.\r\nThis is probably also observable in regular run,when running more than one cursor/object comparison in single test.\r\n\r\n\r\n**Provide version info**\r\nutPLSQL v3.1.8\r\n\r\n**To Reproduce**\r\n```sql\r\nclear screen\r\nset serverout on\r\ndeclare\r\n  l_actual   sys_refcursor;\r\n  l_expected sys_refcursor;\r\nbegin\r\n  open l_actual   for select rownum rn from dual connect by level \u003c 5;\r\n  open l_expected for select rownum rn from dual connect by level = 1;\r\n  ut.expect(l_actual).to_equal(l_expected);\r\n  open l_actual   for select rownum rn from dual connect by level \u003c 3;\r\n  open l_expected for select * from (select rownum rn from dual connect by level \u003c 3) order by 1 desc;\r\n  ut.expect(l_actual).to_equal(l_expected);\r\nend;\r\n/\r\nrollback;\r\n```\r\n\r\nThis produces output:\r\n```\r\nFAILURE\r\n  Actual: refcursor [ count = 4 ] was expected to equal: refcursor [ count = 1 ]\r\n  Diff:\r\n  Rows: [ 3 differences ]\r\n    Row No. 2 - Extra:    \u003cRN\u003e2\u003c/RN\u003e\r\n    Row No. 3 - Extra:    \u003cRN\u003e3\u003c/RN\u003e\r\n    Row No. 4 - Extra:    \u003cRN\u003e4\u003c/RN\u003e\r\n  at \"anonymous block\", line 7\r\nFAILURE\r\n  Actual: refcursor [ count = 2 ] was expected to equal: refcursor [ count = 2 ]\r\n  Diff:\r\n  Rows: [ 2 differences ]\r\n    Row No. 1 - Actual:   \u003cRN\u003e1\u003c/RN\u003e\r\n    Row No. 1 - Expected: \u003cRN\u003e2\u003c/RN\u003e\r\n    Row No. 2 - Actual:   \u003cRN\u003e2\u003c/RN\u003e\r\n    Row No. 2 - Expected: \u003cRN\u003e1\u003c/RN\u003e\r\n    Row No. 2 - Extra:    \u003cRN\u003e2\u003c/RN\u003e\r\n    Row No. 3 - Extra:    \u003cRN\u003e3\u003c/RN\u003e\r\n    Row No. 4 - Extra:    \u003cRN\u003e4\u003c/RN\u003e\r\n  at \"anonymous block\", line 10\r\n```\r\n\r\nHowever adding a commit or rollback in between expectations fixes the issue.\r\n\r\n```sql\r\nclear screen\r\nset serverout on\r\ndeclare\r\n  l_actual   sys_refcursor;\r\n  l_expected sys_refcursor;\r\nbegin\r\n  open l_actual   for select rownum rn from dual connect by level \u003c 5;\r\n  open l_expected for select rownum rn from dual connect by level = 1;\r\n  ut.expect(l_actual).to_equal(l_expected);\r\n  rollback;\r\n  open l_actual   for select rownum rn from dual connect by level \u003c 3;\r\n  open l_expected for select * from (select rownum rn from dual connect by level \u003c 3) order by 1 desc;\r\n  ut.expect(l_actual).to_equal(l_expected);\r\nend;\r\n/\r\nrollback;\r\n```\r\n\r\n```\r\nFAILURE\r\n  Actual: refcursor [ count = 4 ] was expected to equal: refcursor [ count = 1 ]\r\n  Diff:\r\n  Rows: [ 3 differences ]\r\n    Row No. 2 - Extra:    \u003cRN\u003e2\u003c/RN\u003e\r\n    Row No. 3 - Extra:    \u003cRN\u003e3\u003c/RN\u003e\r\n    Row No. 4 - Extra:    \u003cRN\u003e4\u003c/RN\u003e\r\n  at \"anonymous block\", line 7\r\nFAILURE\r\n  Actual: refcursor [ count = 2 ] was expected to equal: refcursor [ count = 2 ]\r\n  Diff:\r\n  Rows: [ 2 differences ]\r\n    Row No. 1 - Actual:   \u003cRN\u003e1\u003c/RN\u003e\r\n    Row No. 1 - Expected: \u003cRN\u003e2\u003c/RN\u003e\r\n    Row No. 2 - Actual:   \u003cRN\u003e2\u003c/RN\u003e\r\n    Row No. 2 - Expected: \u003cRN\u003e1\u003c/RN\u003e\r\n  at \"anonymous block\", line 11\r\n```\r\n\r\n**Expected behavior**\r\nDiff reports should always provide valid outputs.\r\n\r\nTo fix this, we can delete the diff table as part of diff cleanup.\r\n","author":{"url":"https://github.com/jgebal","@type":"Person","name":"jgebal"},"datePublished":"2019-09-22T10:54:05.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":0},"url":"https://github.com/998/utPLSQL/issues/998"}