Java与大数据架构：09-金竹&JEFFChina_Python on Flink & Flink on Zeppelin(1)_compressed.pdf

1、Python on Flink & Flink on Zeppelin ?(?)&?(?) Staff Engineer Alibaba 04/22/2020 About Me 2016.10 Staff Engineer Alibaba Apache Flink 2017.7 Apache IoTDB Committer 2019.6 Apache Flink PMC 2020.1 Apache Flink Committer 2020.2 ALC Beijing Member 2020.4 ASF Member ? ? ? Apache Beam Committer ? Agenda ?

2、Why PyFlink ? Status of PyFlink ? Future of PyFlink ? Flink on Zeppelin PyFlink ? ? ? 85 ? ? ?PyFlink?PyFlink?CDN? ? ? ? ? PyFlink ? ? ? ? ? ? get_execution_environment(.) t_env.register_function(.) t_env.add_python_file(.) t_env.from_path(.) .group_by(.) .select(.) .insert_into(.) t_env.execute(.)

3、Py ? ? ? Py Logging State Data Metric State DataUDF ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? - BeamPy4J Python UDF Java PyFlink - PyFlink ? ? ? ? ? ? ? ? Table API/SQL Python VM Java VM Python UDF Dependency management User-Defined Function Python Lib V1.9 V1.10 Pandas Series/DataFrame Of Gelly API ML Pipeline API Python SQL Client Counters, Gauges, Meters V1.11 UDF Metrics Cython For Performance Tyak