How to turn character cells into double in MATLAB?

I have the following cell array:

load 'qq.mat'

blocks

%100×1 cell array

 %   {'mpta'}
  %  {'mpta'}
   % {'mpta'}
    %{'p'   }
    %{'s0'  }
    %{'s'   }
    %...

I am trying to turn it into double so that I can perform horzcat operations with double objects. Yet, whenever I try to use str2double I get NaNs. Nothing of what I found online works in my case. Can anyone help me with this?

Thanks!

[Walter Roberson]

What numeric values would you like 'mpta' and 's0' to be converted into?

[Armando MAROZZI]

I would like them to remain as they are, but in a format that allows me to use horzcat

Answer 1

[Walter Roberson]

blocks_numeric = double(char(blocks));

You can now horzcat() blocks_numeric and numeric data. The entries will appear as things like

   109   112   116    97
   112    32    32    32
   115    48    32    32

The number of columns will be the same as the maximum length of character vector in blocks, which is 4 in this case.

You can later turn back into text by using

   cellstr(char(TheNumericBlock(:,1:4)))

[Armando MAROZZI]

thanks. However, I need a Nx1 column of names to horzcar with columns of number. That doesn't sort this issue

[Walter Roberson]

What you believe you need cannot be done in MATLAB. Numeric arrays can never have printable characters in them in MATLAB. Or any other programming language that I can think of.

You can use cell() arrays. You can use table() . You can even convert to categorical() and use that in tables to get rid of the apostrophes.

You can convert the double to character.

B = {'mpta'; 'p'; 's0'}

B = 3x1 cell array
    {'mpta'}
    {'p'   }
    {'s0'  }

v = randi(9, 3, 4)

v = 3×4

2 2 5 2 3 6 1 9 3 2 3 2

[B, string(v)]

ans = 3×5 string array
    "mpta"    "2"    "2"    "5"    "2"
    "p"       "3"    "6"    "1"    "9"
    "s0"      "3"    "2"    "3"    "2"

[B, compose("%0.3f", v)]   %to control the exact format of converting the numbers

ans = 3×5 string array
    "mpta"    "2.000"    "2.000"    "5.000"    "2.000"
    "p"       "3.000"    "6.000"    "1.000"    "9.000"
    "s0"      "3.000"    "2.000"    "3.000"    "2.000"

[Armando MAROZZI]

Thanks for the examples, very instructive! But that's very weird to me. You can easily do it in R or Python, creating dataframes with characters and numbers in different columns. Thanks again!

[Walter Roberson]

Python dataframes do not convert the text to numeric form in able to do the equivalent of horzat() them !!

MATLAB has tables, but in tables the text is not horzcat() with the numeric data.

B = {'mpta'; 'p'; 's0'}

B = 3x1 cell array
    {'mpta'}
    {'p'   }
    {'s0'  }

v = randi(9, 3, 4)

v = 3×4

8 9 3 2 7 8 1 8 5 1 1 8

table(B, v)

ans = 3x2 table
       B               v        
    ________    ________________

    {'mpta'}    8    9    3    2
    {'p'   }    7    8    1    8
    {'s0'  }    5    1    1    8

Bc = categorical(B);
table(Bc, v)

ans = 3x2 table
     Bc            v        
    ____    ________________

    mpta    8    9    3    2
    p       7    8    1    8
    s0      5    1    1    8

[Stephen23]

"You can easily do it in R or Python, creating dataframes with characters and numbers in different columns."

Not at all, you are comparing apples with oranges.

R's dataframes are more like MATLAB's tables, they are nothing like a contiguous array of numeric data.

Native Python does not have contiguous arrays of numeric data, for that you need numpy or some similar module.