How can I load a big table in Matlab using textscan

I am trying to load a csv file on matlab using textscan. C is a 1x4 cell array with each element being a 36x1 cell. I try to get the table by using the cell2table function but returns a 1x4 table. I need a 36x4 table but I am not sure why the cell doesnt load as a 36X4 cell in the first place.

% 
f = fopen('test.csv');
formataux = '%s%f%f%s'; 
C =textscan(f, formataux,'Headerlines', 1, 'Delimiter', ',', 'EmptyValue', NaN,'CollectOutput',1); % collect all out put. no need to use struct as I dont have to manipulate each column
fclose(f); 
data = cell2table(C);

[Stephen23]

"C is a 1x4 cell array with each element being a 36x1 cell."

Are you sure? Take a look at cells C{2} and C{3} and tell us what they actually contain (hint: the class is shown in the your screenshot too).

Once you know the real classes, then you will know why cell2table is not doing what you want.

PS: Is there a reason why you are not using readtable (which is likely to make your task a lot easier) ?

[Nikolaos Vasilas]

Hi thanks for you reply. This is just an example Im using to illustrate the task. The actual csv I want to load has 12 million rows and is 1GB long so I thought that text scan should be more efficient. is there a better way?

First column is the stock name (string), 2nd and thrid are open and close price (double ) and the 4th is the stock exchange. I just need to load this in an actual table format, where its column is of different file type.

[Stephen23]

"is there a better way?"

readtable

Or if memory is a resctricting factor, you could use tall arrays:

https://www.mathworks.com/help/matlab/tall-arrays.html

"First column is the stock name (string), 2nd and thrid are open and close price (double ) and the 4th is the stock exchange."

Sure. But did you actually look at the class of the cells 2 and 3 contents, like I told you to? Are they really cell arrays (as your wrote in your question) ? (hint: no)

When you read the cell2table documentation it clearly states "...converts the contents of an m-by-n cell array, C, to an m-by-n table...". Do you have an mxn cell array? (hint: yes, you have a 1xn cell array, and that is the size table that you will get, exactly as the documentation states).

"but I am not sure why the cell doesnt load as a 36X4 cell in the first place."

textscan returns a 1xn cell array, each cell of which can be a column vector or matrix (usually numeric or cell array of char vectors). It does not return an nxm cell array (with n>1).

Either pay particular attention to sizes and classes, or use readtable.