pyflink key_by操作

# -*- coding: utf-8 -*-
from pyflink.datastream import StreamExecutionEnvironment
from pyflink.datastream.functions import  MapFunction, RuntimeContext, KeyedProcessFunction
from abc import ABC, abstractmethod
from pyflink.datastream import StreamExecutionEnvironment
from pyflink.datastream.functions import  MapFunction, RuntimeContext, KeyedProcessFunction
from pyflink.datastream.state import MapStateDescriptor
from pyflink.datastream.connectors.kafka import FlinkKafkaConsumer
from pyflink.common.typeinfo import Types, TypeInformation
from pyflink.datastream.connectors.elasticsearch import Elasticsearch7SinkBuilder, ElasticsearchEmitter, FlushBackoffType
from pyflink.datastream.connectors import DeliveryGuarantee
from pyflink.common.serialization import SimpleStringSchema
import json
import re
from datetime import datetime
from elasticsearch import Elasticsearch
from pyflink.datastream.functions import RuntimeContext, FlatMapFunction

import re
import redis

# 创建 StreamExecutionEnvironment 对象
env = StreamExecutionEnvironment.get_execution_environment()
env.set_parallelism(1)

# 读取文件，创建 DataStream 对象
data_stream = env.read_text_file('/root/pyflink/elink_test.txt')

# 对每行数据添加字符串 'aaaa'

class LogEvent:
    buss_seq = None
    message = None
    
    def __init__(self, bus_seq,message,index_name):
        self.bus_seq = bus_seq
        self.message = message
        self.index_name= index_name

    def to_dict(self):
        return {
            "bus_seq": self.bus_seq,
            "message": self.message,
            "index_name" : self.index_name
        }
class MyMapFunction(FlatMapFunction):
   def open(self, runtime_context: RuntimeContext):
       self.process_id_to_bus_seq = runtime_context.get_map_state(MapStateDescriptor('process_id', Types.STRING(), Types.STRING()))
   def close(self):
       pass

   def flat_map(self,line):
      bus_seq=''
      process_id=''
      if not line.startswith("ES"):
          return 
      if '' in line:
         try:
             pat=re.compile(r"(\d+)")
             bus_seq=pat.search(line).group(1)
             process_id=line.split()[1]
             self.process_id_to_bus_seq.put(process_id, bus_seq)
         except:
             return 
      process_id=line.split()[1]
      if not len(process_id)==6 :
          process_id=line.split()[2]
      bus_seq=self.process_id_to_bus_seq.get(process_id)
      if not bus_seq:
         bus_seq='0'
      #self.r.delete(process_id)
      #log_event = LogEvent(bus_seq.decode('UTF-8'),line)
      #LogEvent['bus_seq']=bus_seq.decode('UTF-8')
      try:
         datetime.now().strftime("%Y-%m-%d")
         index_name='flink-test'+date_str
         log_event=LogEvent(bus_seq,line,index_name)
      except:
          return 
      yield log_event.to_dict()
    
env.add_jars("file:///root/lib/flink-sql-connector-elasticsearch7-3.0.1-1.16.jar")
date_str = datetime.now().strftime("%Y-%m-%d")
es7_sink = Elasticsearch7SinkBuilder() \
    .set_bulk_flush_max_actions(1) \
    .set_emitter(ElasticsearchEmitter.static_index('flink-test2023-06-07')) \
    .set_hosts(['127.0.0.1:9200']) \
    .build()
      

#new_stream = data_stream.map(MyMapFunction()).sink_to(es7_sink)
#new_stream = data_stream.key_by(lambda log: log.split(' ')[1]).flat_map(MyMapFunction(),output_type=Types.MAP(Types.STRING(),Types.STRING())).sink_to(es7_sink)
#new_stream = data_stream.key_by(lambda log: log.split(' ')[1]).flat_map(MyMapFunction(),output_type=Types.MAP(Types.STRING(),Types.STRING()))
new_stream = data_stream.key_by(lambda log: log.split(' ')[1]).flat_map(MyMapFunction(),output_type=Types.MAP(Types.STRING(),Types.STRING())).sink_to(es7_sink)
#new_stream = data_stream.map(MyMapFunction(),output_type=Types.MAP(Types.STRING(),Types.STRING()))
#new_stream = data_stream.map(MyMapFunction(),output_type=Types.MAP(Types.STRING(),Types.STRING()))
# 输出到控制台

# 执行任务
env.execute('Add "bus_seq" to each line')