利用os库批量处理文件或者文件夹

1 批量移动文件

import os
to_path = 'H:/China/LAI'
for year in range(2003, 2021):
    os.chdir(r'H:/China/LAI/%d' % year)
    os.system('move *.tif %s' % to_path)

2 批量删除文件

import os
for i in range(2003,2021):
    for j in range(1,13):
        os.system('cd H:/MYD15A2H/%d/%d/ && del *.hdf && del *.prm' % (i,j))

3 批量修改文件

def rename(startyear=2001,endyear=2003,inputpath=''):
    """save入参数介绍
    循环将各年份中的文件名修改
    比如：tmp2001_1======>tmp200101
    :startyear:开始年份
    :endyear：结束年份+1
    ：inputpath：输入数据目录，比如'H:/test/'
    """
    import os
    for year in range(startyear,endyear):
        path = inputpath+'/'+str(year)
        os.chdir(path)
        fileList=os.listdir(path)   
        n = 0
        for i in fileList:
            #设置旧文件名称（就是路径+文件名）
            oldname=path+'/'+fileList[n]
            print(oldname)
            # 设置新文件名称
            if fileList[n][-5]=='_':
                newname = path+'/'+fileList[n][:-5]+'0'+fileList[n][-4]+'.nc'
            else:
                newname = path+'/'+fileList[n][:-6]+fileList[n][-5:-3]+'.nc'
            print(newname)
            #os.rename(oldname,newname)
            print(oldname,'==============>',newname)
            n = n+1


startyear=2003
endyear=2021
inputpath='H:/China/EVI/nc/'
rename(startyear=startyear,endyear=endyear,inputpath=inputpath)

4 批量给数据添加时间维度

import xarray as xr
import pandas as pd 
import glob
alltimes = []
years = np.arange(2001,2021,1)
months = np.arange(1,12+1,1)
for year in years:
    for month in months:
        newtime = str(year) + '-' + str(month) + '-' + '15'
        #print(newtime)
        alltimes.append(newtime)

def dims_time(data,nu):
    datetime_series=pd.Series(pd.date_range(alltimes[nu],periods=1,freq='d'))
    datetime_series=datetime_series.values.astype("datetime64[D]")
    times=xr.DataArray(datetime_series,[('time',datetime_series)])
    timedata=data.expand_dims(time=times)
    return timedata

rasters2 = glob.glob(r'H:\Climate\pre\*.nc')
for i,raster in enumerate(rasters2):
    #print(raster)
    da=xr.open_dataset(raster)
    var=da.pre*0.1
    newdata=dims_time(var,i)
    #print(i)
    print(raster.split('\\')[3],'========>',newdata.time.values)
    print(np.nanmean(newdata.values))
    newdata.to_netcdf('H:/Climate/test/'+raster.split('\\')[3])