Import .txt data into CELL ARRAY

I have text files with arrays of data:

 { Centroid = [[1 0 1],[0 1 0]] Ages= [10 20 30],Times = [ 3:00, 3:05] }
  ->Exact example, text file includes enclosed brackets '{ }'

I need to import this data into a (cell array?) in MATLAB, so that I can call the various labels and access their data.

[Paolo]

What would the desired cell array look like?

[Philip]

I don't have a particular idea, but some type of table in which I could call chunks of data.

 'Centroid'
 {[0 1 1]}
 {[2 1 3]}

Etc: I'm not super familiar with Tables in MATLAB

Answer 1

[Paolo]

Try something like this:

raw = fileread('mytextfile.txt');
[~,tok]=regexp(raw,'([A-Z][a-z]+).*?(\[.*?(?=[ ,][A-Z}]))','match','tokens');
tokens = [tok{:}];
varnames = tokens(1:2:end);
data = tokens(2:2:end);
cell2table(data,'VariableNames',varnames)

1×3 table

         Centroid              Ages             Times     
    ___________________    ____________    _______________

    '[[1 0 1],[0 1 0]]'    '[10 20 30]'    '[ 3:00, 3:05]'

[Philip]

What is happening in that second row, can you explain?

[~,tok]=regexp(raw,'([A-Z][a-z]+).*?(\[.*?(?=[ ,][A-Z}]))','match','tokens');

[Paolo]

I am happy to explain but before doing so can you just confirm that the output table is what you are after?

[Philip]

That would work fine, yes, as long as I can call those values.

[Paolo]

The second line of code is a regular expression. It can be used to match pattern in strings. In the regular expression, there are two capturing groups. Those are used to extract the name of the variable and the data itself.

Perhaps you can try the regex demo live here ( link)

Please note that I am assuming that this pattern is consistent across your entire text file.

[Philip]

That's intense, thanks, I will have to play with it and lookup documentation to understand each component of this.

[Philip]

Question on output: Can you now call specific parts from the table? How would you do that: example

answer = table('Ages',array[0])
...
(answer =  10)

Would it be better/easier to use textscan? Especially for text files that are just one big array of numbers?

Thanks!