Github user fhueske commented on a diff in the pull request: https://github.com/apache/flink/pull/5218#discussion_r159136001 --- Diff: flink-java/src/test/java/org/apache/flink/api/java/io/RowCsvInputFormatTest.java --- @@ -362,61 +365,95 @@ public void readStringFieldsWithTrailingDelimiters() throws Exception { @Test public void testTailingEmptyFields() throws Exception { - String fileContent = "abc|-def|-ghijk\n" + - "abc|-def|-\n" + - "abc|-|-\n" + - "|-|-|-\n" + - "|-|-\n" + - "abc|-def\n"; - FileInputSplit split = createTempFile(fileContent); - - TypeInformation[] fieldTypes = new TypeInformation[]{ - BasicTypeInfo.STRING_TYPE_INFO, - BasicTypeInfo.STRING_TYPE_INFO, - BasicTypeInfo.STRING_TYPE_INFO}; - - RowCsvInputFormat format = new RowCsvInputFormat(PATH, fieldTypes, "\n", "|"); - format.setFieldDelimiter("|-"); - format.configure(new Configuration()); - format.open(split); - - Row result = new Row(3); - - result = format.nextRecord(result); - assertNotNull(result); - assertEquals("abc", result.getField(0)); - assertEquals("def", result.getField(1)); - assertEquals("ghijk", result.getField(2)); + List<Tuple4<TypeInformation, String, String, String>> dataList = new java.util.ArrayList<>(); + + // test String + dataList.add(new Tuple4<>(BasicTypeInfo.STRING_TYPE_INFO, "bdc", "bdc", "")); + // test BigInt + dataList.add(new Tuple4<>(BasicTypeInfo.BIG_INT_TYPE_INFO, --- End diff -- I think we can extend this test but should not exhaustively test all types. It's the responsibility of the different `FieldParser` to implement the empty field logic correctly. Hence, this should be tested in the field parser tests and not here.
---