from scipy.io import loadmat
随便从下面文件里读取一个：
m = loadmat(‘H_BETA.mat‘) # 读出来的 m 是一个dict（字典）数据结构
读出来的m内容：
m：
{‘__header__‘: b‘MATLAB 5.0 MAT-file, Platform: GLNXA64, Created on: Mon Aug 5 17:14:09 2019‘, ‘__version__‘: ‘1.0‘, ‘__globals__‘: [], ‘H_BETA‘: array([[ 0.68508148, 0.36764355, 0.73505849, ..., 0.27600164, 0.67968929, 0.70506438], [ 0.74920812, 1.10949748, 0.47506305, ..., 0.32871445, 0.61247345, 1.06948844], [ 0.83311522, 1.06321302, 0.97364609, ..., 0.85837753, 0.96296771, 1.46095171], ..., [ nan, nan, nan, ..., nan, nan, -9.04648469], [ nan, nan, nan, ..., nan, nan, nan], [ nan, nan, nan, ..., nan,
In [29]: m.keys() Out[29]: dict_keys([‘__header__‘, ‘__version__‘, ‘__globals__‘, ‘H_BETA‘])
.mat 文件里的数据结构是 dict ，所以取值要按照 key:value 的形式：
In [30]: m[‘H_BETA‘] Out[30]: array([[ 0.68508148, 0.36764355, 0.73505849, ..., 0.27600164, 0.67968929, 0.70506438], [ 0.74920812, 1.10949748, 0.47506305, ..., 0.32871445, 0.61247345, 1.06948844], [ 0.83311522, 1.06321302, 0.97364609, ..., 0.85837753, 0.96296771, 1.46095171], ..., [ nan, nan, nan, ..., nan, nan, -9.04648469], [ nan, nan, nan, ..., nan, nan, nan], [ nan, nan, nan, ..., nan, nan, nan]]) In [31]: type(m[‘H_BETA‘]) Out[31]: numpy.ndarray
上面读出来的数据是 ndarray 类型，为了方便数据的展示，我们可以将其转换为，pandas的DataFrame：
In [32]: import pandas as pd In [33]: df = pd.DataFrame(m[‘H_BETA‘]) In [34]: df.head()
Out[34]: 0 1 2 3 4 5 6 7 8 9 10 0 0.685081 0.367644 0.735058 0.085046 0.104332 0.560731 0.350219 0.758185 0.303823 0.114022 0.452877 1 0.749208 1.109497 0.475063 0.896100 1.117772 0.611356 0.662669 0.603077 0.863930 0.756870 0.725808 2 0.833115 1.063213 0.973646 0.935061 0.631670 0.916800 0.662993 0.543231 0.671558 1.027954 0.526402 3 0.488906 0.932741 0.956622 0.573116 0.893764 0.987304 0.380807 1.211157 0.550213 0.898408 1.153289 4 0.440694 0.503209 0.509693 0.477054 0.344717 -0.054662 1.124213 0.344906 0.612898 0.217625 -0.129715 [5 rows x 2111 columns]
如此，数据就比较规整了，是保存成文件，还是做其他处理，就by yourself啦！
原文：https://www.cnblogs.com/bigtreei/p/11307872.html