String with csv in matlab

Hi I have a string in matlab A = "2011-08-18,00:10:00,600,6.2,5.4, " etc... Can i make this Csv-string into data? (I have a lot of these from an earlier bad import)

Thank you! Fredrik

[Jan]

It depends on what "data" means exactly. Please explain the type of the input (string, cell string, test file, MAT file?) and the wanted output explicitely.

[Fredrik]

First of all, i have a cell array called Text with a lot of strings, that look like the one above. Text{1} and Text{2} is

'2011-08-18,00:10:00,600,1.56188,3.33684,4.328568,2.03864' and

'2011-08-18,00:20:00,600,7.26988,7.23684,0.328568,8.03864'

What I would like to do is eventually to get the values from these strings into a matrice with NewText(1)= '2011-08-18, 00:10:00'

and Data (1,1)= 600 for example.

So first my thought was to export this out from matlab in some way and then import it back with csvread for example, but would be more efficient to do it all inside matlab!

Answer 1

[TAB]

Text={'2011-08-18,00:10:00,600,1.56188,3.33684,4.328568,2.03864';'2011-08-18,00:20:00,600,7.26988,7.23684,0.328568,8.03864'};

TextLen=length(Text);
NewText=cell(TextLen,2);
Data=zeros(TextLen,5);
for x=1:length(Text)
    SplitData=textscan(Text{x},'%s','delimiter',',');
    NewText{x}={SplitData{1}{1},SplitData{1}{2}};
    Data=(str2double({SplitData{1}{3:end}}));
end

[Fredrik]

Perfect, thanx alot for your time and a great answer!

Answer 2

[Fredrik]

Do you have any idea on how to increase the speed of this? This way of importing takes about 1 second per file I import, and since it is about 50 of them, it would be nice for some more speed :) For example, is it possible to use the textscan without going through a for loop, more like a vector operation?

[TAB]

Try *cellfun*

See doc cellfun