File Exchange

image thumbnail


version 1.0 (2.12 KB) by

reads Excel's files stored in .csv or .xls file formats and stores results as a struct

9 Ratings



View License

CSV2STRUCT reads Excel's files stored in .xls or .csv file formats and stores results as a struct.
The Excel file is assumed to have a single header row. The output struct will have a field for each column and the field name will be based on the column name read from the header.
Unlike csvread, csv2struct is able to read files with both text and number fields and store data as fields of a struct. Likely works on Windows machines only.

See also:
MATLAB's csvread and xlsread functions
xml_read from my xml_io_tools which creates struct out of xml files

Comments and Ratings (14)


Haldun (view profile)

Hi Jaroslaw,

When i call this funciton within another script i get "Expression or statement incorrect-possibly unbalanced" error. The error refers to Line 33 Column 3, which corresponds to tilde operator. Do you have an idea ?

Jason Baxter

I need to transfer data structures from Matlab to C, I was hoping this would do it but this doesn't work if the structures have more than one level.

For example:

>> test.a1.a2.a4=1;
>> test.a1.a2.a2=10;
>> test.a11.a2.a1=1;
>> test.b=[12;42;12];
>> test.c={'safds','sdfs'};

Runs but produces nonsense. Please fix or list the limitations in the description and help entry.


Andrea (view profile)

I keep getting the following error message when trying to run the code...

Invalid field name: '@'.

Error in csv2struct_TEST (line 82)
Out.(name) = num(:,iNum);

Any suggestions for debugging? Am I missing something?


Andrew (view profile)

chang hsiung

can this be used in Mac ?


Thai (view profile)

Great code.


Awesome script! CSV only works for me if i have a header with more than one character.

Jeff Barton

Jeff Barton (view profile)

Just what I was looking for. Thank you.


shlomix (view profile)

it didn't work well for me when I passed a csv file which size exceeded 2^16.


Pete (view profile)

@Ged: "***If your system does not have Excel for Windows installed, or MATLAB cannot access the COM server, xlsread operates in basic mode. In this mode, xlsread only reads XLS files.***"

Yeah this is unfortunate - would seem to mean that this script will only work on windows machines. Not the author's fault, but down-rated accordingly =/

Michael Chan

Michael Chan (view profile)

It's really useful.
just a minor error.

line 78: name = char('A'-1+c);
instead of
name = char('A'-1+i);


Alfred (view profile)

Ged Ridgway

Really useful submission!

Works nicely on .xls, but doesn't seem to work on .csv for me... In particular, xlsread (which csv2struct relies on even for .csv) doesn't seem happy with .csv files at all; I'm not sure if this is a platform issue (xlsread under Windows can be more advanced than xlsread under Linux, which I'm currently using) or a version issue (I'm currently using R2006a). Any ideas? (It's easy enough to convert csv to xls, and then use it, but csv is more widely compatible than xls for other software...)


Pete (view profile)

Have only just started playing with this script, but so far it seems to be another truly excellent utility from Jaroslaw Tuszynski.

Does everything it says. Columns with text come out as cells. Columns with numbers come out as matrices. And all of them are stored in a structure such that you can reference them by the name of the column header.

This seems to me far superior to anything else I've found either here on FEX, or built in to Matlab (though I'm only running R2007a at the moment, so I can't comment on newer releases).

Will post an update if I come across any problems further down the line, but on present experience I highly recommend this script!



Updated syntax and removed no longer needed code for reading CSV files with more than 2^16 rows. Excel no longer has that limitation.

MATLAB Release
MATLAB 7.7 (R2008b)

Inspired: Struct array conversion

Download apps, toolboxes, and other File Exchange content using Add-On Explorer in MATLAB.

» Watch video