python+pandas气象数据处理

最近帮同事处理一批气象数据，统计某省多个地市50年的的平均降雨量，excle折腾半天，还可能卡死，python几十行代码即可，不得不佩服pandas库的强大。

"""
  @Author  : ISR
  @contact: [email protected]
  @Software: PyCharm
  @FileName: get_meteorology.py
  @Time    : 2020/5/18 9:46
  @desc:
"""
import pandas as pd
from pandas.core.frame import DataFrame
import numpy as np

def main(in_path, out_path, star,end):

    # 读取气象数据
    df = pd.read_csv(in_path, header=0, index_col=0, engine='python')

    data = np.zeros((len(citys), 2))

    for i in range(len(citys)):
		
		# 进度条
        pst = int(100 * (i + 1) / len(citys))
        print('\r{0}{1}{2}%'.format(("计算中".split("\\")[-1]), '▉' * int(pst / 10), (pst)), end='')

        dataframe = df[(df['站名'] == citys[i]) & (df['year'] >= star) & (df['year'] <= end)]

        # 秋季 6月到9月
        autumn = df[(df['站名'] == citys[i]) & (df['year'] >= star) & (df['year'] <= end) & (df['mouth'] < 10) & (df['mouth']  >=6)]

        # 春季 10月到次年6月（这里利用pandas去重函数drop_duplicates，实际是为了取余）
        dataframe = dataframe.append(autumn)
        spring = dataframe.drop_duplicates(keep=False)
		
		# 最后一列为降雨量
        autumn = (autumn.values)[:, -1]
        spring = (spring.values)[:, -1]
		
		# 计算年均值
        autumn_mean = (np.sum(autumn[np.where(autumn > 0)]))/(end-star+1)
        spring_mean = (np.sum(spring[np.where(spring > 0)]))/(end-star+1)

        data[i, 0] = autumn_mean
        data[i, 1] = spring_mean

    meteorology = DataFrame(data, columns=["autumn","spring"], index=citys)

    meteorology.to_csv(out_path)

if __name__ == '__main__':

    citys = [“石家庄”，“天津”] # 站名

    star = 1990 # 起始年份
    end = 2019 # 结束年份

    in_path =r"C:\Users\marshmallow\Desktop\04_气象数据\in.csv" # 输入
    out_path = r"C:\Users\marshmallow\Desktop04_气象数据\out.csv"# 输出

    main(in_path, out_path, star, end)

1、输入数据结构图（降雨数据倒数第一列）。

2、将一年分成了冬、求两季。
秋季作物-当年6月到当年9月
春季作物–当年10月到次年5月