How to properly apply cellfun in this case - text spliting

I've got a table like this and my task is to separate the parameter name from a value:

t =

    10×1 table

            Var1      
      ________________

      'Pa     0.749µm'
      'Pq     0.800µm'
      'Pz     1.458µm'
      'Pp     0.748µm'
      'Pv     0.710µm'
      'Pt     4.914µm'
      'Psk    -0.321' 
      'Pkµ    1.726'  
      'Pc     E_0010' 
      'PSm    E_0010'

So first I came up with the code to identify all of the spaces in the cells and then fins the first and last blank in each row:

separator = ' ';
BlanksIdx = strfind(t{:,1},separator); % Indexes of all blanks in the rows
Blank1st  = cellfun(@min, BlanksIdx);
BlankLast = cellfun(@max, BlanksIdx);

With the output for Blank1st being:

Blank1st =

       3
       3
       3
       3
       3
       3
       4
       4
       3
       4

At this point I wanted to use cellfun() again to carry out a separation:

A = cellfun(@(x) x(1:Blank1st-1),t{:,1},'UniformOutput',false)

However, this always truncates the text after 2nd character:

A =

    10×1 cell array

      {'Pa'}
      {'Pq'}
      {'Pz'}
      {'Pp'}
      {'Pv'}
      {'Pt'}
      {'Ps'}
      {'Pk'}
      {'Pc'}
      {'PS'}

How do I make this code to work so it separates the name of the parameter correctly?

Answer 1

[Stephen23]

It is simpler and more efficient to use a regular expression:

C = regexp(t.Var1,'\S+','match');
C = vertcat(C{:})

Giving the two columns in one cell array:

>> C{:,1}
ans = Pa
ans = Pq
ans = Pz
ans = Pp
ans = Pv
ans = Pt
ans = Psk
ans = Pkµ
ans = Pc
ans = PSm
>> C{:,2}
ans = 0.749µm
ans = 0.800µm
ans = 1.458µm
ans = 0.748µm
ans = 0.710µm
ans = 4.914µm
ans = -0.321
ans = 1.726
ans = E_0010
ans = E_0010

Answer 2

[Rik]

You need to have the index available as a separate input:

A = cellfun(@(x,ind) x(1:ind-1),t{:,1},num2cell(Blank1st),'UniformOutput',false)

Answer 3

[Pawel Jastrzebski]

Rik Wisselink, Stephen Cobeldick many thanks to you both for the suggestions. I'll go forward with the regexp solution as it seems cleaner.