How can I use Python pandas in MATLAB?

Looking at this tutorial (https://uk.mathworks.com/help/matlab/matlab_external/call-user-defined-custom-module.html), it appears easy to use a Python module in MATLAB.

I'd like to execute the following Python code snippet in MATLAB:

import pandas
tbl = pandas.read_csv('xxxxx.csv')

However, the following does not work in MATLAB.

tbl = py.pandas.read_csv('xxxxx.csv')

Any suggestions?

Answer 1

[Kouichi C. Nakamura]

pyversion returns Python installation outside of Anaconda.

>> pyversion
       version: '2.7'
    executable: 'C:\Python27\python.EXE'
       library: 'C:\WINDOWS\system32\python27.dll'
          home: 'C:\Python27'
      isloaded: 1

So I needed to change the reference to the Anaconda version of Python.

After relaunching MATLAB, I used pyversion again to change the Python path. Then py.pandas.read_csv() worked!

>> pyversion 'C:\Users\xxxxxxxxx\AppData\Local\Continuum\anaconda2\python.EXE'
>> pyversion
       version: '2.7'
    executable: 'C:\Users\xxxxxxxxx\AppData\Local\Continuum\anaconda2\python.EXE'
       library: 'C:\Users\xxxxxxxxx\AppData\Local\Continuum\anaconda2\python27.dll'
          home: 'C:\Users\xxxxxxxxx\AppData\Local\Continuum\anaconda2'
      isloaded: 1
>> py.pandas.read_csv('xxxxxxxxx.csv')
ans = 
  Python DataFrame with properties:

          T: [1×1 py.pandas.core.frame.DataFrame]
         at: [1×1 py.pandas.core.indexing._AtIndexer]
       axes: [1×2 py.list]
     blocks: [1×1 py.dict]
    columns: [1×1 py.pandas.core.indexes.base.Index]
      empty: 0
        iat: [1×1 py.pandas.core.indexing._iAtIndexer]
       iloc: [1×1 py.pandas.core.indexing._iLocIndexer]
      index: [1×1 py.pandas.core.indexes.range.RangeIndex]
    is_copy: [1×1 py.NoneType]
         ix: [1×1 py.pandas.core.indexing._IXIndexer]
        loc: [1×1 py.pandas.core.indexing._LocIndexer]
       ndim: 2
      shape: [1×2 py.tuple]
       size: 120
      style: [1×1 py.pandas.io.formats.style.Styler]
     values: [1×1 py.numpy.ndarray]
       ...
    0  ...
    1  ...
    2  ...
    3  ...
    4  ...
    5  ...
    6  ...
    7  ...
    8  ...
    9  ...
    10 ...
    11 ...
    12 ...
    13 ...
    14 ...
    
    [15 rows x 8 columns]      

[Dominik Mattioli]

You didn't need to point python to your anaconda environment too? Particularly, where the environment in which pandas is installed?

[Brett Swanson]

It probably worked because the default Anaconda environment ("base") has a huge number of modules (including pandas) installed.

Answer 2

[Artem Lenskiy]

Check this submission https://au.mathworks.com/matlabcentral/fileexchange/111770-pandastomatlab

It implements two funcions:

• df2t - that converts Pandas DataFrame to Matlab Table
• t2df - that converts Matlab Table to Pandas DataFrame.

The examples are shown in test.mlx. Here is one example

Name            = {["Roger", "Sanchez"];
                   ["Paul", "Johnson"];
                   ["Lisa", "Li"];
                   ["Don", "Diaz"];
                   ["Havana ", "Brown"]};
Age             = [38;43;38;40;49];
Smoker          = logical([1;0;1;0;1]);
Height          = [71;69;64;67;64];
Weight          = [176;163;131;133;119];
BloodPressure   = [124 93; 109 77; 125 83; 117 75; 122 80];
T = table(Name,Age,Smoker,Height,Weight,BloodPressure);
T.BMI           = (T.Weight * 0.453592)./(T.Height * 0.0254).^2;
df              = t2df(T);                              % Convert Table to Python Pandas
% Sample from the dataframe
df_sampled      = df.sample(int64(10), replace=true);   % Call DataFrame functions
table_sampled   = df2t(df_sampled)                      % Convert DataFrame back to Table