Using "find" in cell inside cell with different data types

Hello,

my data is 1x253 cell, inside each are 253 further cell matrices with different sizes. An example of one of the cells is:

    'COP PE Equity'  NaN      NaN  'COR PE Equity'          NaN  NaN  NaN      NaN
      735452  6332  'NoValue'   633  '#N/A Invalid Security'  NaN  'NoValue'0       NaN
      735475  2232  'NoValue'   223    NaN                  NaN   NaN   NaN     NaN
      735545  304.1  'NoValue'   304.1 NaN                  NaN   NaN   NaN     NaN
      NaN  NaN  NaN      NaN   NaN                          NaN   NaN   NaN
      NaN  NaN  NaN      NaN    NaN                          NaN   NaN   NaN
      NaN  NaN  NaN      NaN    NaN                          NaN   NaN  NaN
      NaN  NaN  NaN      NaN    NaN                          NaN   NaN  NaN
      NaN  NaN  NaN      NaN    NaN                          NaN   NaN  NaN
      NaN  NaN  NaN      NaN    NaN                          NaN   NaN  NaN
      NaN  NaN  NaN      NaN    NaN                          NaN   NaN  NaN

I would like to implement the following:

for i=1:4:3497
    if strcmp(data{1,a}{2,i},'#N/A Invalid Security') ==0 && strcmp(data{1,a}{2,i},'#N/A N/A') ==0 && strcmp(data{1,a}{2,i},'Date') ==0 && strcmp(data{1,a}{2,i},'TURNOVER') ==0 
    for j=1:size(date,1)
        index = find(date(j,1)==data{1,a}(:,i));
            if isempty(index)
           final{1,a}(j,ceil(i/4)) = nan;
            else
               final{1,a}(j,ceil(i/4)) = data{1,a}{index,i+4};
            end
     end
    end
end

So, I would like to go in each cell matrix inside cell "data", search for each element from "date" vector, and use the fourth column value for the orresponding row (index) in a new cell "final".

When I run this code, I get error message

"Function 'subsindex' is not defined for values of class 'cell'.".

When I try to use:

find(date(j,1)==cell2mat(data{a}(:,i))),

I get error

"Error using cell2mat (line 45) All contents of the input cell array must be of the same data type."

Could you please advise me on how to proceed?

Thanks a lot.

[Walter Roberson]

Note: the expression

if strcmp(data{1,a}{2,i},'#N/A Invalid Security') ==0 && strcmp(data{1,a}{2,i},'#N/A N/A') ==0 && strcmp(data{1,a}{2,i},'Date') ==0 && strcmp(data{1,a}{2,i},'TURNOVER') ==0

can be simplified to

if ~ismember(data{1,a}{2,i}, {'#N/A Invalid Security', '#N/A N/A', 'Date', 'TURNOVER'})

but you would probably be better using

if ischar(data{1,a}{2,i}) && ~ismember(data{1,a}{2,i}, {'#N/A Invalid Security', '#N/A N/A', 'Date', 'TURNOVER'})

[Ekaterina Serikova]

Thanks. Any suggestions on how to make "find" function work?

[Guillaume]

In my opinion, you would be better off changing the way you imported that data. This would make your current task and any other task where you have to search particular values much easier (even possibly trivial).

You seem to have headers mixed with data. The headers should be separated (in another cell array, or even better using tables which are much easier to use than cell arrays)

You have a mix of char arrays and numbers in the same column. What is the difference between NaN and 'NoValue'. Probably none, so these 'NoValue' could be converted to NaN.

Once the type of each column is clearly defined, it would be much easier to search them. Using table(s) instead of cell arrays would make it even easier.

[Ekaterina Serikova]

Dear Guillaume, if I'm importing 255 tables, is there a way to locate them into one variable or I have to create 255 different variables for each of them?

[Walter Roberson]

You can create cell arrays of tables.

[Guillaume]

... or you could have one table where each column is a cell array of column, depending on what's better for later processing.

A cell array of tables would be easier to generate but possibly harder to search.

Alternatively, you could have just one big table with an additional column indicating which file/cell array the data came from. This would make searching/filtering almost trivial.