Modify Strings in cell array according to pattern

I have to cell arrays A & B

A = ['C1 H2 O1 N1 S1';
    [];
'C5 H2 O0 N1 S0';
'C2 H6 O0 N1 S1';
'C2 H6 O2 N1 S0';
[];
'C3 H10 O1 N1 S0';
'C6 H5 O2 N2 S1';
'C8 H9 O4 N0 S0';
[];
'C9 H13 O3 N0 S0';
'C10 H17 O2 N0 S0';
'C11 H21 O1 N0 S0';
'C12 H25 O0 N0 S0'];

and

B = ['C5H9O2+'
'C3H5O4'
'C3H7O5']

and I would like to have B in the same pattern as A so that (C* H* O* N* S*)

and deleting the +

B[1] = 'C5 H9 O2 NO S0'

kinda tried with regexprep() but didn't get anything by now.....

Answer 1

[Allen]

Let me know how this works for you.

B = {'C3H7O';'C2H7O2';'C2H7O2+';'C4H5O';'C4H5O';'C3H3O2';'';'C4H7O';'C3H5O2';'C3H7O2';'C2H5O3';...
    'C2H5O3';'C2H7O3';'';'';'C4H5O2';'C4H5O2';'C4H5O2';'C4H7O2'};
I = {'C0','H0','O0','N0','S0'};

B = regexprep(B,{'\W','(?![CHONS]\d+)[CHONS]'},''); % Removes all but letter/number pairs
B(cellfun(@isempty,B)) = join(I,''); % Replaces empty cells with I

% Adds missing letter/number pairs to each cell of B and joins with no spacing
expr = {'(C\d+)','(H\d+)','(O\d+)','(N\d+)','(S\d+)'};
for i=1:size(B,1)
    B(i) = join([B{i},I(cellfun(@isempty,regexp(B{i},expr)))],''); 
end

% Adds a single space between letter/number pairs
B = regexprep(B,'(C\d+)(H\d+)(O\d+)(N\d+)(S\d+)','$1 $2 $3 $4 $5');

[Martin Muehlegger]

You'r a genius thanks.

[Allen]

You're welcome.

One thing to be aware of is that if the order of CHONS matter and they are not in this order in your original B variable, then you will need to tweak this code some.

[Martin Muehlegger]

And here we are,.... I have to tweak the code,

My new B is

B = ['C9H19NO2', 'C9H20O3', 'C6H12O6', 'C11H20O2', 'C10H19NO2', 'C10H21NO2', ...
    'C10H20O3', 'C11H19NO2', 'C11H18O3', 'C12H22O2', 'C11H21NO2', 'C11H20O3']

So the order has changed to CHNOS and i would like to have it in the order as before CHONS and filled up with letter/number pair to each cell of B

PS: found a problem in the first code, when in B some letter without a number was found it shouldn't be zero e.g.

'C3H7O' = C3 H7 O0 N0 S0

but instead it should be

'C3H7O' = C3 H7 O1 N0 S0

because there is one 'O'

thanks

Answer 2

[Allen]

I am sure that there is a much more elegant method using just regular expressions, but the following should give you what you are looking for if your input cell-array is similar to the example provided.

B = {'C5H9O2+'
     'C3H5O4'
     'C3H7O5'};
I = {'C','H','O','N','S'};

% Looks for 'C', 'H', 'O', 'N',and 'S' and when not present in a cell-array element, append the
% missing charater followed by a '0' to the cell-array element.
for i=1:length(I)
    idx = ~contains(B,I{i});
    B(idx) = append(B(idx),I{i},'0');
end
% Use regular expression to replace all non-alphanumeric or underscore characters, then find
% matching tokens consisting of 'C', 'H', 'O', 'N', and 'S' each followed by any number of digits
% and recombine them as a single space-delimited string.
regexprep(B,{'\W','(C\d*)(H\d*)(O\d*)(N\d*)(S\d*)'},{'','$1 $2 $3 $4 $5'})

[Martin Muehlegger]

Thanks Allen, appreciate your help! Got an error according to append();

%Error using append (line 38)
%Wrong number of input arguments for obsolete matrix-based syntax.

Tried to fix it with:

B(idx) = [B(idx); I{i}; '0']; instead of %B(idx) = append(B(idx),I{i},'0');

But regexprep gives me an error as well

% In an assignment  A(:) = B, the number of elements in A and B must be the same.

Here is a bigger set of my Cell B.

B = ['C3H7O';
'C2H7O2';
'C2H7O2+'
'C4H5O';
'C4H5O';
'C3H3O2';
''; % the empty shouldn't make any trouble 
'C4H7O';
'C3H5O2';
'C3H7O2';
'C2H5O3';
'C2H5O3';
'C2H7O3';
'';
'';
'C4H5O2';
'C4H5O2';
'C4H5O2';
'C4H7O2']

[Allen]

What version of MATLAB are you using? Also, you appear to be defining B as a non-cell-array instead of a cell-array. That might also be part of the problem.

[Martin Muehlegger]

Matlab 2017b

whos B
  Name        Size            Bytes  Class    Attributes

  B         177x1             21946  cell