energy-data-trans/tmp_file/张崾先筛选20241210.py

import datetime
import multiprocessing
import os
import sys

import pandas as pd

sys.path.insert(0, os.path.abspath(__file__).split("tmp_file")[0])

from utils.file.trans_methods import read_file_to_df, read_excel_files

# read_dir = r'/data/download/collection_data/1进行中/张崾先风电场-陕西-华电/清理数据/点检表以外测点儿-20241209'

# save_dir = r'/data/download/collection_data/1进行中/张崾先风电场-陕西-华电/清理数据/变桨-20241210'

# user_cols = ['Time', '机舱外温度', '桨叶角度A', '桨叶角度B', '桨叶角度C',
#              '轴1电机电流', '轴2电机电流', '轴3电机电流',
#              '轴1电机温度', '轴2电机温度', '轴3电机温度']


read_dir = r'/data/download/collection_data/1进行中/张崾先风电场-陕西-华电/清理数据/点检表以外测点儿-20241210'

save_dir = r'/data/download/collection_data/1进行中/张崾先风电场-陕西-华电/清理数据/偏航-20241210'

user_cols = ['Time', '实际风速', '偏航误差', '电缆扭角', '激活偏航解缆阀','激活顺时针偏航','激活逆时针偏航']

os.makedirs(save_dir, exist_ok=True)


def read_and_save(file_path, read_dir, save_dir):
    begin = datetime.datetime.now()
    df = read_file_to_df(file_path, read_cols=user_cols)
    df['Time'] = pd.to_datetime(df['Time'], errors='coerce')
    df.sort_values(by=['Time'], inplace=True)
    df.to_csv(os.path.join(save_dir, os.path.basename(file_path)), index=False, encoding='utf8')
    print(os.path.basename(file_path), '耗时:', (datetime.datetime.now() - begin))


if __name__ == '__main__':
    begin = datetime.datetime.now()
    all_files = read_excel_files(read_dir)

    with multiprocessing.Pool(16) as pool:
        pool.starmap(read_and_save, [(file, read_dir, save_dir) for file in all_files])

    print('总耗时:', (datetime.datetime.now() - begin))