Spliting Cell Into Multiple Columns

I have been having issues separating my cell data into multiple columns. But I seem to be encountering problems with trying to split my 1 column cell data into 4 columns.

I have attached my code below:

 %%Data Import:
 %  Open File:
    FileID     = fopen('TCOMP244.TXT');
 %  Read File
    TCOMP_CELL = textscan(FileID,'%s','Delimiter','\n');
 %  Close File:
    fclose(FileID);
 %  Data Extraction:
    DATA_CELL=TCOMP_CELL{1,1};

 %%Data Extraction:
 %  Iteration Initialization:
    [k, j, l, m, o, p, r] = deal(1, 1, 1, 1, 1, 1, 1);

 % Cell Initialization:
    Primary_Chain      = cell((ceil((length(DATA_CELL)-182)./89).*20),1);
    Primary_Am_Chain   = cell((ceil((length(DATA_CELL)-211)./89).*15),1);
    Secondary_Am_Chain = cell((ceil((length(DATA_CELL)-235)./89).*15),1);

 %%Primary Fission Chain:
    for i=182:89:length(DATA_CELL)
        for k=0:19
            Primary_Chain{j,1} = DATA_CELL{i+k};
            j = j+1;
        end
    end

 %%Principal Americium Chain:
    for l=211:89:length(DATA_CELL)
        for n=0:14
            Primary_Am_Chain{m,1} = DATA_CELL{l+n};
            m = m+1;
        end
    end

 %%Secondary Americium Chain:
    for o=235:89:length(DATA_CELL)
        for q=0:14
            Secondary_Am_Chain{p,1} = DATA_CELL{o+q};
            p = p+1;
        end
    end

This leads to following output:

What should I do to separate the cell into multiple columns.

*My intent is to take the latter 3 columns and convert them into a floating point matrix for use in a separate calculation.

Thanks in advance.

[jonas]

You can probably avoid a lot of headache if you import the data properly, in this step:

TCOMP_CELL = textscan(FileID,'%s','Delimiter','\n');

if you attach the data it's easier to help you. In the meantime you can try readtable which often figures out how to import tabular data without additional inputs.

[Quang Phung]

Hi,

Here's the data if you would like to take a look at it. Unfortunately, I have a rather limited amount of experience with importing and processing data with MATLAB, so using 'textscan' seemed to be the most convenient for me at the time.

[jonas]

I'll take a look. textscan is probably the most flexible input option but also one of the more difficult ones to master. I'd consider myself novice :)

Can you explain exactly what you want to store from this text? It seems to have one big table and some smaller ones.

EDIT: I'll have a look when I wake up tomorrow if no one helped you by that time. This is probably the most complex table I've ever tried to import.

[Quang Phung]

Yes, so the program I used to generate the data calculates the current mass and other nuclide properties at specific times. These are then exported into a text file (see Image 1).

This is the data for one chain (out of three different chains) at two minutes.

The next iteration of that chain occurs (89 lines down) at line 271. Only at a new time of 12 hours (See Image 2).

This continues about 13 times.

The loop I currently have coalesces each chain iteration into a single cell. You can see this in image 3.

Once done, my intention is to split that cell from 1 column into 4 columns. Remove the first column (nuclide names) and then convert the last 3 columns from strings into floating point.

Edit: Thanks for your assistance. I agree, a lot of this is junk data too. :/

[Stephen23]

@Quang Phung: exactly which data do you need to import from that file? Can you please show us which values you need.

[Quang Phung]

Hi,

As I explained previously, the program I use to generate this data calculates various properties of nuclides at discrete points in time (determined by the user). In this case, it calculates the data from 3 different chains. Image 1 details such chain (The primary chain), which always starts at line 182. The other two chains start at lines, 211 and 235 (which can be seen in my nested loops). Each chain then restarts at a new time 89 lines down (regardless of the number of time steps the user selects).

The chain data is calculated at each time step (in this case 13 times, but the number is dependent on the number of time steps the user imports).

The current code I have reads the data from TCOMP244.txt using textscan and (presumably converts each line into a string).

The three nested loop then reads the data from each chain and coalesces each individual chain into a composite chain. E.g. Primary_Chain contains all the data from the primary chain (20 nuclides by 13 intervals = 260).

My intent is to split the chain data from a 260x1 cell into a 260x4 cell. Where each column is the various nuclide data and the nuclide name. I intend to delete the first column (nuclide names) and convert the remaining 3 columns from a cell into a matrix. My issue is splitting the cell into 4 columns.

The data I need can be seen in Image 1 and Image 2 (that I posted in response to Jonas). Image 3 is the cell data I extracted from the primary chain (Image 1 and Image 2). You can see that the data repeats itself starting at line 21 of the cell.

Answer 1

[Robert U]

Hello Quang Phung,

for your data "Primary_Chain" you could use the following code fragment in order to

• seperate the columns inside cell
• delete 1st column
• convert to array of double

    % separate columns
    Test = cellfun(@(cIn) strsplit(cIn,' ')',Primary_Chain,'UniformOutput',false);
    Test = [Test{:}]';

    % delete first column
    Test(:,1) = [];

    % convert to double
    TestOut = cellfun(@str2num,Test);

Kind regards,

Robert

[Quang Phung]

Hi Robert,

Thank you so much, the cell splitting function works perfectly. Could you explain how you created your code fragment. I'd like to be able to understand how this works for future reference.

Thanks!!!

[Robert U]

• strplit() with delimiter ' '(space) is applied to each cell of cell array Primary_Chain
• output is a cell array again that contains the strings split to cells
• recover sub-cells via calling all sub-cells and collect in an array (to not loose structure mind the transpose()-command twice, once within cellfun() and next after data has been collected
• delete first column straight forward
• apply str2num() to each cell of Test, output will be numeric array of doubles