Core,API: Set 503: added_snapshot_id as required

[Fokko]

503: added_snapshot_id field should be written as a required field, but currently it is written as optional.

As the reference implementation should produce metadata that is as close to the spec as possible.

For reading, this isn't a problem with the current Java implementation as it will still read optional fields as required, but only thrown an error when it encounters a null value.

[Fokko]

See

iceberg/core/src/main/java/org/apache/iceberg/SnapshotParser.java

Lines 179 to 218 in a52afdc

	/**
	* The main purpose of this class is to lazily retrieve the path from a v1 Snapshot that has
	* manifest lists
	*/
	private static class DummyFileIO implements FileIO {
	@Override
	public InputFile newInputFile(String path) {
	return new InputFile() {
	@Override
	public long getLength() {
	throw new UnsupportedOperationException();
	}
	@Override
	public SeekableInputStream newStream() {
	throw new UnsupportedOperationException();
	}
	@Override
	public String location() {
	return path;
	}
	@Override
	public boolean exists() {
	return true;
	}
	};
	}
	@Override
	public OutputFile newOutputFile(String path) {
	throw new UnsupportedOperationException();
	}
	@Override
	public void deleteFile(String path) {
	throw new UnsupportedOperationException();
	}
	}

[rdblue]

Minor: it looks like this is done quite a bit. What about making it easier either with a test helper or a better builder?

[rdblue]

Overall looks good. I'm trying to think if there was a reason why the spec requires snapshot ID but we didn't write it that way. I can't think of what might have caused this. Good catch!

[Fokko]

Great suggestion @rdblue, thanks for the review!