错误地读取使用Python的NetCDF和HDF文件

Question

我有以下代码，可以让我打开 .nc 或 .h5 文件，并且可以打开其内容。

import gdal
import glob
import os
import numpy as np
import pandas


def nch5read(fn, ncmaskscale=True):
    import netCDF4 as nc
    import h5py
    from pyhdf.SD import SD, SDC

    d = {'VARS':{}, 'ATTRS':{}}
    if os.path.splitext(fn)[1] == '.nc':
        f = nc.Dataset(fn)
        f.set_auto_maskandscale(ncmaskscale)
        for i in f.ncattrs():
            try:
                d['ATTRS'][i] = f.getncattr(i)
            except:
                pass

        for i in f.variables:
            try:
                d['VARS'][i] = {'DATA':np.squeeze(f.variables[i][:]),
                'ATTRS':{j:f.variables[i].getncattr(j) for j in f.variables
                }}
            except:
                pass

    if os.path.splitext(fn)[1]== '.h5':
        f = h5py.File(fn)
        d['ATTRS']=dict(f.attrs)

        for i in f:
            try:
                d['VARS'][i] = {'DATA':np.squeeze(f[i][()]),
                'ATTRS':dict(f[i].attrs)}
            except:
                pass

    return d

fn = "cons_dom.nc"

d = nch5read(fn)
print(d)

我正在使用从该网站获取的示例生活用水文件 cons_dom.nc https://zenodo.org/record/897933#.YkH_SefMKUn 但在运行代码时，我得到一个空白结果 {'VARS': {}, 'ATTRS': {}}

另外，我尝试在代码中添加 .hdf 部分

import gdal
import glob
import os
import numpy as np
import pandas


def nch5read(fn, ncmaskscale=True):
    import netCDF4 as nc
    import h5py
    from pyhdf.SD import SD, SDC

    d = {'VARS':{}, 'ATTRS':{}}
    if os.path.splitext(fn)[1] == '.nc':
        f = nc.Dataset(fn)
        f.set_auto_maskandscale(ncmaskscale)
        for i in f.ncattrs():
            try:
                d['ATTRS'][i] = f.getncattr(i)
            except:
                pass

        for i in f.variables:
            try:
                d['VARS'][i] = {'DATA':np.squeeze(f.variables[i][:]),
                'ATTRS':{j:f.variables[i].getncattr(j) for j in f.variables
                }}
            except:
                pass

    if os.path.splitext(fn)[1]== '.h5':
        f = h5py.File(fn)
        d['ATTRS']=dict(f.attrs)

        for i in f:
            try:
                d['VARS'][i] = {'DATA':np.squeeze(f[i][()]),
                'ATTRS':dict(f[i].attrs)}
            except:
                pass
# Segment added to read .hdf files
    if os.path.splitext(fn)[1]== '.hdf':
        f = SD(fn, SDC.READ)
        f.set_auto_maskandscale(ncmaskscale)
        for i in f.variables:
            try:
                d['VARS'][i] = {'DATA':np.squeeze(f.variables[i][:]),
                'ATTRS':{j:f.variables[i].getncattr(j) for j in f.variables}}
            except:
                pass
    return d

fn = "MCD12Q1.A2016001.h10v05.006.2018149125004.hdf"

d = nch5read(fn)
print(d)

但是当我尝试跑步示例 MODIS 文件，MCD12Q1.A2016001.h10v05.006.2018149125004.hdf 我收到错误

File "C:\Users\Jay_Oliver\Anaconda3\envs\rasterio\lib\site-packages\pyhdf\SD.py", line 3222, in _getattr
    index = a.index()
  File "C:\Users\Jay_Oliver\Anaconda3\envs\rasterio\lib\site-packages\pyhdf\SD.py", line 1193, in index
    _checkErr('find', self._index, 'illegal attribute name')
  File "C:\Users\Jay_Oliver\Anaconda3\envs\rasterio\lib\site-packages\pyhdf\error.py", line 23, in _checkErr
    raise HDF4Error(err)
pyhdf.error.HDF4Error: find (59): Invalid arguments to routine

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "nc_example.py", line 63, in <module>
    d = nch5read(fn)
  File "nc_example.py", line 52, in nch5read
    f.set_auto_maskandscale(ncmaskscale)
  File "C:\Users\Jay_Oliver\Anaconda3\envs\rasterio\lib\site-packages\pyhdf\SD.py", line 1446, in __getattr__
    return _getattr(self, name)
  File "C:\Users\Jay_Oliver\Anaconda3\envs\rasterio\lib\site-packages\pyhdf\SD.py", line 3224, in _getattr
    raise AttributeError("attribute not found")
AttributeError: attribute not found