Extracting numbers from cell using regex

Hi,

I have the following cell:

C = {'2/6';'78/6';'982/47';'11/6'}

I would like to extract all numbers from before / and place them in a new matrix, and all numbers from after / and place them in another matrix.

The result should be something like this:

A=[2;78;982;11] B=[6;6;47;6]

How can i do this? I guess using regex would be the easiest way, but how to write it so it extracts it as explained?

Thank you for your help.

Answer 1

[Walter Roberson]

C = {'2/6';'78/6';'982/47';'11/6'};

One way:

temp = cell2mat(cellfun(@str2double,regexp(C, '(\d+)/(\d+)', 'tokens','once'),'uniform',0));
A = temp(:,1);
B = temp(:,2);

Another way:

temp = cell2mat(regexp(C, '(?<A>\d+)/(?<B>\d+)', 'names'));
A = str2double({temp.A}.');
B = str2double({temp.B}.');

Another way:

temp = cell2mat(cellfun(@str2double,regexp(C,'/','split'),'uniform',0));
A = temp(:,1);
B = temp(:,2);

Another way:

A = str2double(regexp(C,'\d+(?=/)','match','once'));
B = str2double(regexp(C,'(?<=/)\d+','match','once'));

Answer 2

[Jos (10584)]

You can use cellfun to scan every string:

C = {'2/6';'78/6';'982/47';'11/6'}

V = cellfun(@(c) sscanf(c,'%f/%f'), C, 'un', 0) ;
AB = cat(2,V{:})
A = AB(1,:)
B = AB(2,:)

[kalili]

I am sorry, I was a bit wrong when writing the question. The data is of type cell, not string: C = {2/6;78/6;982/47;11/6}

[Walter Roberson]

Jos and I have assumed your C was what you posted before, a cell array of character vectors. Our answers work with the C you had posted earlier. But if your C is a cell array of numeric values:

>> C = {2/6;78/6;982/47;11/6}
C =
  4×1 cell array
    {[0.333333333333333]}
    {[               13]}
    {[ 20.8936170212766]}
    {[ 1.83333333333333]}

then it is not possible to recover the original numerators and denominators: you can just get approximations of the reduced fractions:

[A,B] = rat(cell2mat(C))

[Jos (10584)]

@kalili ... sigh ...

Well, how did you get, for instance, the 2 and the 6 in cell 1 in the first place? Maybe you can skip that step?

[kalili]

Hm, the data was obtained by importing an excel sheet using [~,~,C]=xlsread(filename).

[Walter Roberson]

What shows up for

class(C{1})

When you use the third output of xlsread(), then anything that looks like a plain number will be converted to numeric but anything else will be left alone. Approximately speaking, it is like

   temp = read cells as text
   temp_numeric = str2double(temp);
   mask = ~isnan(temp_numeric);
   temp(mask) = num2cell(temp_numeric(mask))

so '2/6' should have been left alone as text because str2double() if it would be nan. Note that str2num() is not used for this purpose: str2num() would evaluate the text as an expression, getting the 0.333<etc> but that is not going to be used.