Python将方法放置到线程池中，多线程执行；方法执行效率翻倍提升！

import multiprocessing as mp
import traceback
import pandas as pd


def process_df_chunk(df_chunk: object, iterrows_map: dict) -> object:
    """
    对块数据进行批处理，通过vocab字典拿到对应的函数对象及参数
    :param df_chunk: dataFrame
    :param iterrows_map:{obj:['','','',..]}
    :return: dataFrame
    """
    for f in iterrows_map:
        for i, row in df_chunk.iterrows():
            if row['language_type'] in ['zh', 'en']:
                df_chunk.at[i, iterrows_map[f][0]], df_chunk.at[i, iterrows_map[f][1]] = f(
                    row[iterrows_map[f][2]], iterrows_map[f][3]
                )
    return df_chunk

def process_df(df: object, iterrows_map: dict, num_processes=6) -> object:
    """
    分块处理，完成后重写组装数据
    :param df:
    :param iterrows_map: {obj:['','','',..]}
    :param num_processes:
    :return: dataFrame
    """
    df_chunks = [df[i:i + num_processes] for i in range(0, df.shape[0], num_processes)]
    with mp.Pool(processes=num_processes) as pool:
        df_list = pool.starmap(process_df_chunk, [(chunk, iterrows_map) for chunk in df_chunks])
    return pd.concat(df_list)