group可以当作一个dataset一样处理,两个类拥有大多数相同的函数,在一个group中也能继续创建子group。group中的维度可以与全局维度一起使用,如果维度名称相同,则在创建变量的时候会使用group中的维度。
import netCDF4 as nc
# 创建数据
dataset = nc.Dataset(r'H:\testgroup.nc',mode='w')
# 创建group
group = dataset.createGroup('1')
dataset.close()
import netCDF4 as nc
import numpy as np
dataset = nc.Dataset(r'H:\testgroup.nc',mode='w')
group = dataset.createGroup('1')
# 创建维度和变量
group.createDimension('lat',10)
group.createDimension('lon',10)
group.createVariable('lat',np.float32,('lat'))
group.createVariable('lon',np.float32,('lon'))
group['lat'][:] = np.arange(10)
group['lon'][:] = np.arange(10)
dataset.close()
import netCDF4 as nc
import numpy as np
dataset = nc.Dataset(r'H:\testgroup.nc',mode='w')
# 创建全局维度
dataset.createDimension('lat',11)
dataset.createDimension('lon',11)
dataset.createVariable('lat',np.float32,('lat'))
dataset.createVariable('lon',np.float32,('lon'))
dataset.variables['lat'][:] = np.arange(11)
dataset.variables['lon'][:] = np.arange(11)
# 创建group
group = dataset.createGroup('1')
group.createDimension('lon',10)
group.createVariable('lat',np.float32,('lat'))
group.createVariable('lon',np.float32,('lon'))
group['lat'][:] = np.arange(11)
# 将10改成11会报错,说明这里的lon使用的是group中的维度
group['lon'][:] = np.arange(10)
dataset.close()
根据我的需求,我需要创建上千个group,它们使用同一组维度,因此将维度信息存储在全局中,只存储一次,每一组中存放不同时期的数据。
import netCDF4 as nc
import numpy as np
dataset = nc.Dataset(r'H:\testgroup.nc',mode='w')
dataset.createDimension('lat',10)
dataset.createDimension('lon',10)
dataset.createVariable('lat',np.float32,('lat'))
dataset.createVariable('lon',np.float32,('lon'))
dataset.variables['lat'][:] = np.arange(10)
dataset.variables['lon'][:] = np.arange(10)
for i in range(10):
group = dataset.createGroup(str(i))
group.createVariable('water_temp',np.float32,('lat','lon'))
group['water_temp'][:] = np.random.rand(10,10)
dataset.close()
import netCDF4 as nc
import numpy as np
dataset = nc.Dataset(r'H:\testgroup.nc')
for group_name in dataset.groups:
group = dataset.groups[group_name]
for var_name in group.variables.keys():
print(group.name,group[var_name][:])
dataset.close()