[Bug]: [Java] Row equality not working as expected when ByteBuffer used as Map key

[ahmedabu98]

What happened?

I was writing a test that checked for the equality of Beam Rows with Maps that have ByteBuffer keys. Row's .equals() method doesn't account for this case.

To reproduce:

Schema schema = Schema.of(Schema.Field.of("bytesMap", Schema.FieldType.map(Schema.FieldType.BYTES, Schema.FieldType.STRING)));

ByteBuffer bytes1 = ByteBuffer.wrap("a".getBytes(StandardCharsets.UTF_8));
Map<ByteBuffer, String> map1 = new HashMap<>();
map1.put(bytes1, "abc");
Row row1 = Row.withSchema(schema).withFieldValue("bytesMap", map1).build();

ByteBuffer bytes2 = ByteBuffer.wrap("a".getBytes(StandardCharsets.UTF_8));
Map<ByteBuffer, String> map2 = new HashMap<>();
map2.put(bytes2, "abc");
Row row2 = Row.withSchema(schema).withFieldValue("bytesMap", map2).build();

System.out.println("map equals: " + map1.equals(map2)); // true
System.out.println(row1.equals(row2)); // false

Although ByteBuffer is used as the map key, Row stores the byte array value instead. There are known issues when using byte[] type as Map key (equality is determined by array identity, so only the exact same byte[] key can be used to retrieve a value. ie, a new identical byte[] will not retrieve any values. Read here for more details).

Issue Priority

Priority: 3 (minor)

Issue Components

• Component: Python SDK
• Component: Java SDK
• Component: Go SDK
• Component: Typescript SDK
• Component: IO connector
• Component: Beam examples
• Component: Beam playground
• Component: Beam katas
• Component: Website
• Component: Spark Runner
• Component: Flink Runner
• Component: Samza Runner
• Component: Twister2 Runner
• Component: Hazelcast Jet Runner
• Component: Google Cloud Dataflow Runner

[Amar3tto]

.take-issue