python实现pyflink的table自定义函数

from pyflink.table import EnvironmentSettings, TableEnvironment
from pyflink.table import EnvironmentSettings, TableEnvironment, DataTypes

from pyflink.table.udf import TableFunction, udtf
from pyflink.table.expressions import call
from pyflink.table.udf import udf

# 1. 创建 TableEnvironment
env_settings = EnvironmentSettings.new_instance().in_streaming_mode().use_blink_planner().build()
table_env = TableEnvironment.create(env_settings)

# 2. 创建 source 表
table_env.execute_sql("""
    CREATE TABLE datagen (
        i INT,
        j INT,
        m VARCHAR 
    ) WITH (
              'connector' = 'kafka',
              'topic' = 'flink_udftest1',
              'properties.bootstrap.servers' = '1.21:9092',
              'properties.group.id' = 'test_PrintUdf',
              'scan.startup.mode' = 'latest-offset',
              'format' = 'json'
            )
""")


# @udf(input_types=[DataTypes.INT(), DataTypes.INT()], result_type=DataTypes.INT())
# def add(i, j):
#     return i + j
# table_env.register_function("add", add)



@udtf(result_types=[DataTypes.STRING(), DataTypes.STRING()])
def splitTest(s: str):
    splits = s.split("#")
    yield splits[0], splits[1]
# 注册 Python 表值函数
#split = udtf(splitTest(), result_types=[DataTypes.STRING(), DataTypes.STRING()])

table_env.register_function("splitStr", splitTest)
# 3. 创建 sink 表i INT,
#         j INT,m VARCHAR
table_env.execute_sql("""
    CREATE TABLE print (
        m1 VARCHAR ,
        m2 VARCHAR 

    ) WITH (
        'connector' = 'print'
    )
""")

# 4. 查询 source 表，同时执行计算
# 通过 Table API 创建一张表：
source_table = table_env.from_path("datagen")

# 或者通过 SQL 查询语句创建一张表：
# source_table = table_env.sql_query("SELECT * FROM datagen")

#source_table.execute()

#table_env.create_temporary_function("split", udtf(splitTest(), result_types=[DataTypes.STRING(), DataTypes.STRING()]))
#table_env.sql_query("SELECT m1, m2 FROM datagen, LATERAL TABLE(splitStr(m)) as T(m1, m2)")
#table_env.sql_query("SELECT a, word, length FROM MyTable LEFT JOIN LATERAL TABLE(split(a)) as T(word, length) ON TRUE")

table_env.sql_update("insert into print SELECT m1, m2 FROM datagen, LATERAL TABLE(splitStr(m)) as T(m1, m2)")
table_env.execute("job_sum")