Issue fix for CSV loading with header and skip header not parsing well. (#10398)

extensions-core/lookups-cached-global/src/main/java/org/apache/druid/query/lookup/namespace/UriExtractionNamespace.java

@@ -216,6 +216,10 @@ public class UriExtractionNamespace implements ExtractionNamespace
     public Map<String, String> parseToMap(String input)
     {
       final Map<String, Object> inner = delegate.parseToMap(input);
+      if (null == inner) {
+        // Skip null or missing values, treat them as if there were no row at all.
+        return ImmutableMap.of();
+      }
       final String k = Preconditions.checkNotNull(
           inner.get(key),
           "Key column [%s] missing data in line [%s]",
@@ -296,9 +300,10 @@ public class UriExtractionNamespace implements ExtractionNamespace
           this.valueColumn,
           Arrays.toString(columns.toArray())
       );
-
+      CSVParser csvParser = new CSVParser(null, columns, hasHeaderRow, skipHeaderRows);
+      csvParser.startFileFromBeginning();
       this.parser = new DelegateParser(
-          new CSVParser(null, columns, hasHeaderRow, skipHeaderRows),
+          csvParser,
           this.keyColumn,
           this.valueColumn
       );
@@ -401,6 +406,7 @@ public class UriExtractionNamespace implements ExtractionNamespace
           hasHeaderRow,
           skipHeaderRows
       );
+      delegate.startFileFromBeginning();
       Preconditions.checkArgument(
           !(Strings.isNullOrEmpty(keyColumn) ^ Strings.isNullOrEmpty(valueColumn)),
           "Must specify both `keyColumn` and `valueColumn` or neither `keyColumn` nor `valueColumn`"

extensions-core/lookups-cached-global/src/test/java/org/apache/druid/query/lookup/namespace/UriExtractionNamespaceTest.java

@@ -96,7 +96,25 @@ public class UriExtractionNamespaceTest
     );
     Assert.assertEquals(ImmutableMap.of("B", "C"), parser.getParser().parseToMap("A,B,C"));
   }
-
+  @Test
+  public void testCSVWithHeader()
+  {
+    UriExtractionNamespace.CSVFlatDataParser parser = new UriExtractionNamespace.CSVFlatDataParser(
+        ImmutableList.of("col1", "col2", "col3"),
+        "col2",
+        "col3",
+        true,
+        1
+    );
+    // parser return empyt list as the 1 row header need to be skipped.
+    Assert.assertEquals(ImmutableMap.of(), parser.getParser().parseToMap("row to skip "));
+    //Header also need to be skipped.
+    Assert.assertEquals(ImmutableMap.of(), parser.getParser().parseToMap("col1,col2,col3"));
+    // test the header is parsed
+    Assert.assertEquals(ImmutableList.of("col1", "col2", "col3"), parser.getParser().getFieldNames());
+    // The third row will parse to data
+    Assert.assertEquals(ImmutableMap.of("val2", "val3"), parser.getParser().parseToMap("val1,val2,val3"));
+  }
   @Test(expected = IllegalArgumentException.class)
   public void testBadCSV()
   {
@@ -146,6 +164,26 @@ public class UriExtractionNamespaceTest
     );
     Assert.assertEquals(ImmutableMap.of("B", "C"), parser.getParser().parseToMap("A\\u0001B\\u0001C"));
   }
+  @Test
+  public void testWithHeaderAndListDelimiterTSV()
+  {
+    UriExtractionNamespace.TSVFlatDataParser parser = new UriExtractionNamespace.TSVFlatDataParser(
+        ImmutableList.of("col1", "col2", "col3"),
+        "\\u0001",
+        "\\u0002", "col2",
+        "col3",
+        true,
+        1
+    );
+    // skipping one row
+    Assert.assertEquals(ImmutableMap.of(), parser.getParser().parseToMap("Skipping some rows"));
+    // skip the header as well
+    Assert.assertEquals(ImmutableMap.of(), parser.getParser().parseToMap("col1\\u0001col2\\u0001col3"));
+    // test if the headers are parsed well.
+    Assert.assertEquals(ImmutableList.of("col1", "col2", "col3"), parser.getParser().getFieldNames());
+    // test if the data row is parsed correctly
+    Assert.assertEquals(ImmutableMap.of("B", "C"), parser.getParser().parseToMap("A\\u0001B\\u0001C"));
+  }
 
   @Test(expected = IllegalArgumentException.class)
   public void testBadTSV()