Fix String Frame Readers to read String Arrays correctly (#16885)

While writing to a frame, String arrays are written by setting the multivalue byte. But while reading, it was hardcoded to false.
apache · Sep 10, 2024 · c7c3307 · c7c3307
1 parent 72fbaf2
commit c7c3307
Show file tree

Hide file tree

Showing 7 changed files with 603 additions and 274 deletions.
diff --git a/processing/src/main/java/org/apache/druid/frame/field/StringFieldReader.java b/processing/src/main/java/org/apache/druid/frame/field/StringFieldReader.java
@@ -480,9 +480,8 @@ public int numRows()
         public boolean isNull(int rowNum)
         {
           final long fieldPosition = coach.computeFieldPosition(rowNum);
-          byte[] nullBytes = new byte[3];
-          dataRegion.getByteArray(fieldPosition, nullBytes, 0, 3);
-          return Arrays.equals(nullBytes, EXPECTED_BYTES_FOR_NULL);
+          return dataRegion.getByte(fieldPosition) == StringFieldWriter.NULL_ROW
+                 && dataRegion.getByte(fieldPosition + 1) == StringFieldWriter.ROW_TERMINATOR;
         }
 
         @Override

diff --git a/processing/src/main/java/org/apache/druid/frame/read/columnar/FrameColumnReaders.java b/processing/src/main/java/org/apache/druid/frame/read/columnar/FrameColumnReaders.java
@@ -51,15 +51,15 @@ public static FrameColumnReader create(
         return new DoubleFrameColumnReader(columnNumber);
 
       case STRING:
-        return new StringFrameColumnReader(columnNumber, false);
+        return new StringFrameColumnReader(columnNumber);
 
       case COMPLEX:
         return new ComplexFrameColumnReader(columnNumber);
 
       case ARRAY:
         switch (columnType.getElementType().getType()) {
           case STRING:
-            return new StringFrameColumnReader(columnNumber, true);
+            return new StringArrayFrameColumnReader(columnNumber);
           case LONG:
             return new LongArrayFrameColumnReader(columnNumber);
           case FLOAT: