Performance Issue using cell() to convert Python data types to Matlab

Question

Daniel 2024-8-1

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2142146-performance-issue-using-cell-to-convert-python-data-types-to-matlab

回答： Yair Altman 2024-8-4

I have a Python list of 50,000 elements in Matlab, and using the cell() function to convert it to Matlab cell array is taking over an hour, while on my coworkers machine it takes less than a minute. We are both using Python 3.9, Matlab 2021b, and my VM is v4cpu 16G RAM. Also, when trying to pause or stop the execution of cell() Matlab freezes, and has to be forcefully shut down.

Note - The code below executed instantly in this website, but takes over an hour on Matlab 2021b installed on my local.

pyrun("py_list = list(range(50000))");
cellArray = cell(pyrun('py_list', 'py_list'));

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Stephen23 2024-8-1

Contact technical support: https://www.mathworks.com/support/contact_us.html

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Yair Altman 2024-8-4

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2142146-performance-issue-using-cell-to-convert-python-data-types-to-matlab#answer_1494336

If you are certain that the python list only contains numeric values (as in your example code), then using double() instead of cell() should be significantly faster. This results in a numeric array; if you need a cell array for any reason (although it's usually better to stick with numeric arrays for such cases), then you could always wrap the results in a call to num2cell(). The combination of num2cell(double(...)) could be 5-10x faster than cell(...), and without the num2cell() even faster.