xlsread for many distinct columns that aren't side-by-side in the SS

Hi,

I know this is probably simple, but I'm not finding what I need when searching here.

I want to use xlsread to import many distinct columns that are not side-by-side in the spreadsheet. How can I do that? What am I missing? :)

My normal use of xlsread would be like this if I need columns A-D:

[File,Path] = uigetfile(blah, blah, blah);
File = strcat(Path,File);  
DataVariable = xlsread(File,1,A:A:D:D);

How would I change this if I wanted, say, columns A and columns D-F? I've tried many different things, but have suffered death by commas, semicolons, and error messages.

Any help is appreciated.

[John]

Bumping this back up... :)

[Jan]

Please, John, do not bump your question after a few hours without addibng new information. The voluntary helpers post answers whene ever they have them and find the time to help. Reading a question again is a waste of time then.

The posted code is not working. Did you forget the quotes around A:A:D:D? But even then this is not correct.

Answer 1

[Jan]

Import the block at first:

[File,Path]  = uigetfile(blah, blah, blah);
File         = fullfile(Path,File);   % Smarter than STRCAT
DataVariable = xlsread(File,1,'A1:F231');

Now remove the not needed data from the output DataVariable.

This is not nice, but it works reliably.

[John]

I actually tried that way, but ML told me there wasn't enough storage to import that much info. It's trying to import millions of cells.

[Image Analyst]

Does your xlsread call specify a range? You might have to process your data in smaller chunks.

Answer 2

[Image Analyst]

Have you tried readtable() - it's new in R2013b and reads the Excel workbook into a table which is a lot more convenient than that mess of cell arrays xlsread() spits out. I've kissed xlsread() goodbye. You will too after you use readtable().

[Image Analyst]

Here's an example:

% Read in Book1 which has numbers in column A
% nothing at all in column B, and letters in column C.
t = readtable('book1.xlsx');
% I had column B be empty so it's all nans.
% Delete that column
t(:,2)=[];
% Now t has 2 columns like we want (columns A and C from Excel);

[Image Analyst]

You can also use xlsread if you have to

[numbers, strings, raw] = xlsread(filename);
colA = raw(:,1);
colDF = raw(:,4:6);

The only reason you'd do it this way instead of readtable is if you need to run this on an old version of MATLAB (before R2013b).

[John]

I can try that, but I'm fearful I'll run into what I ran into in the above comment under Jan Simon's answer (that ML will tell me there's not enough storage to complete it).

[John]

I have R2013a, so no readtable for me :(. Looks like it would have worked better too.

Answer 3

[Sean de Wolski]

Use an automation server to read in only the chunks you need:

http://www.mathworks.com/help/releases/R2013b/matlab/matlab_external/example-reading-excel-spreadsheet-data.html

[John]

Actually can't view that without entering a license. The license associated with my ML was already used for a MathWorks acocunt by the person before me at work.

[Sean de Wolski]

If you contact our customer service department they can straighten that out.

Otherwise just search for "Excel Automation" in the MATLAB doc; it should be the first result.

[John]

So, Excel Automation is working nicely. However, there's an issue I'm running into.

It requires that you define which sheet you want to use by way of using the sheet name. The sheet name, in my case, is different for every excel file, but does contain part of the file name. The file name is so long, it only allows part of the file name to be placed in the sheet name. My file names I'm using contain many bits of info separated by underscores.

XXX_###_ABC_###_YYY_blah_blah_blah

But, in the sheet name, it cuts off the filename at the same underscore each time.

Can I tell ML to take the filename up until, say, the fifth underscore, and use that as the sheet name?

Or is there a generic way to tell it to define the sheet name as whatever the sheetname actually is? Something similar to the Activesheet command in VBA?

Answer 4

[Nick Haddad]

This issue is a known bug in MATLAB and has been addressed in the following bug report:

• xlsread might return an error for large spreadsheets

The bug report has a workaround which you can install for MATLAB R2013a through R2014b.