技術標籤：pythonpandas

利用Pandas的groupby求站點氣溫月均值

# -*- coding: utf-8 -*-
"""
@title:
@author: Richard Chen
@time:2020-12-28 20:46
"""
import pandas as pd
import numpy as np
import os

path = 'D:/BaiduNetdiskDownload/china_isd_lite_2020/china_isd_lite_2020/'
fileList = os.listdir(path)
 

if path.exists(path+'Month'):
    print('資料夾已建立')
else:
    os.makedirs(path+'Month')
getdata_loop = 0
for i, name in enumerate(fileList):
    oldname = path + fileList[i]
    newname = path + fileList[i]+'.csv'
    os.rename(oldname,newname)
    site = name.split('-')[0]
    dataheader = ['Year', 'Month'
 
, 'Day', 'Hour', 'Temp', 'DPT', 'SLP', 'SD', 'WSR', 'SC', 'LPDD1', 'LPDD6']
    df = pd.read_csv(newname,header = None,names=dataheader,delim_whitespace=True)
    df = df[~df['Temp'].isin([-9999])]
    year = name.split('-')[2]
    result = df.groupby('Month')['Temp'].agg(TempMean='mean',TempMax='max'
 
,TempMin=np.min)/10
    result['Year'] = year
    result['Site'] = site
    if getdata_loop == 0:
        result.to_csv(path+'Month/'+'all.csv',mode='a')
    else:
        result.to_csv(path+'Month/'+'all.csv',mode='a',header=False)
    getdata_loop += 1

location = pd.read_csv('D:\\BaiduNetdiskDownload\\isd-history.csv')
location = location.loc[:,['USAF','LAT','LON']]
location.columns = ['Site','Lat','Lon']
df1 = pd.read_csv(path+'Month/'+'all.csv')
df1['Site'] = df1['Site'].astype(str)
result = pd.merge(df1, location, how='left', on=['Site', 'Site'])
result.to_csv(path+'Month/'+'Temp'+year+'.csv')