要进行Apache Flink Stateful Functions Python与Java性能比较，可以按照以下步骤进行：

1. 安装Apache Flink和Stateful Functions：首先需要安装Apache Flink和Stateful Functions的Python和Java SDK。请参考Apache Flink的官方文档进行安装。

2. 编写Python和Java代码：编写一个简单的Stateful Functions应用程序，可以使用Python和Java编写相同的功能。例如，可以编写一个计数器应用程序，计算收到的事件数量。

Python示例代码：

import time
from statefun import StatefulFunctionsApp, RequestReplyHandler
from statefun import StatefulFunctionsApp, RequestReplyHandler, IntType, from_protobuf, to_protobuf
from statefun.request_reply_protocol import TypedValue
from statefun.statefun import ValueSpec
from statefun import StatefulFunctionsApp, RequestReplyHandler
from statefun import ValueSpec, kinesis_egress_record
from statefun import StatefulFunctionsApp, RequestReplyHandler
from statefun import kafka_egress_record

app = StatefulFunctionsApp("my_app")

@app.functions.bind("com.example.counter")
def counter(context, message):
    state = context.state('count').unpack(IntType).value_or(0)
    state += 1
    context.state('count').pack(state)
    context.send("com.example.sink", message)
    return

@app.functions.bind("com.example.sink")
def sink(context, message):
    print("Received message: ", message)
    return

handler = RequestReplyHandler(app)

while True:
    raw_input = input("Enter a message: ")
    message = TypedValue()
    message.value = bytes(raw_input, 'utf-8')
    handler.handle_sync("com.example.counter", "user-id", message)

Java示例代码：

import org.apache.flink.statefun.sdk.annotations.*;
import org.apache.flink.statefun.sdk.state.*;

public class CounterFunction implements StatefulFunction {

  @Override
  public void invoke(Context context, Object input) {
    ValueState countState = context.states().getValueState("count", Integer.class);
    Integer count = countState.getOrDefault(0);
    countState.set(count + 1);
    context.send(EgressIdentifier.of("com.example.sink", "out"), input);
  }
}

public class SinkFunction implements StatefulFunction {

  @Override
  public void invoke(Context context, Object input) {
    System.out.println("Received message: " + input);
  }
}

public class MyApp implements StatefulFunctionModule {

  @Override
  public void configure(StatefulFunctionModuleConfig.Builder builder) {
    builder.bindFunctionProvider(CounterFunction.class, unused -> new CounterFunction());
    builder.bindFunctionProvider(SinkFunction.class, unused -> new SinkFunction());
    builder.bindEgress("com.example.sink", unused -> new EgressBuilder().withOutputType(String.class).build());
  }
}

public class Main {

  public static void main(String[] args) throws Exception {
    StatefulFunctionsUniverse universe = StatefulFunctionsUniverse.withoutMetrics();
    universe.addModule(new MyApp());
    StatefulFunctionsRuntime runtime = StatefulFunctionsRuntime.start(universe);
    runtime.awaitTermination();
  }
}

3. 运行Python和Java代码：分别运行Python和Java代码，可以通过命令行或IDE来运行。

4. 进行性能测试：使用Apache Bench等工具进行性能测试，比较Python和Java版本的性能差异。可以测试不同并发数和消息数量下的性能表现。

这样就可以通过以上步骤进行Apache Flink Stateful Functions Python与Java性能比较，并得到性能差异的结果。请注意，由于Python是解释性语言，相比于Java，通常会有更大的性能损失。因此，在某些情况下，Java版本的性能可能会更好。