将 matlab.object 转换为 pandas dataframe

Question

I have used this code to run a matlab script inside python, the script is the file'matlab_to_python.m' and it has 1 output: a matlab table called 'matlab_table'.

import matlab.engine
eng = matlab.engine.start_matlab()

mytable = eng.matlab_to_python()

当我运行它时，mytable 的类型是：

In [28]: mytable
Out[28]: <matlab.object at 0x1ebe007fb70>

我希望能够在 python 中将此表视为 pandas dataframe，知道如何完成吗？

谢谢！

Answer 1

尽管似乎尚不支持将表从 MATLAB 直接传输到 Python 中，但您可以利用字符转换来弥补差异。

For example, you might convert the table into json format (1xN char), which could then be turned into a DataFrame using the pandas.DataFrame.read_json function.

import matlab.engine
import pandas as pd
eng = matlab.engine.start_matlab()

# keep the table in MATLAB
eng.workspace['mytable'] = eng.matlab_to_python()

# pass a string representation to Python
jsontable = eng.eval('jsonencode(mytable)')

# convert back into a table
mytable = pd.read(jsontable)

An alternate approach could be to write from matlab to a csv file, which you can then read into Python via pandas.DataFrame.read_csv