Number of .mat files into single .mat file

Hi,

I've 10 .mat files. I want to overlap all 10 .mat files into one (single) mat file.

For ex: my filename is data1, data2,......,data10.mat.

Can someone help me to solve this problem???

Thanks in advance

[Walter Roberson]

Are the variable names different between the files?

If not then should the new files use new variable names? Should it use a struct in which the top level files are the file names? If the variables are the same names in the files, should they be merged together ?

[Kishore]

Sorry, All 10 .mat files variables having same names in the files. Now, I want to overlap all 10 files into a single .mat file.

[Matt J]

@Kishore If they contain the same variable names, what does it mean to "overlap" all 10 files? A single .mat file cannot contain 2 variables with the same name.

Answer 1

[Matt J]

It depends what they contain and how you want the variables joined within the final .mat file. One way:

filenames="data"+(1:10)+".mat";

for i=1:numel(filenames)
    
    S(i)=load(filenames(i));
    
end

save("JoinedData.mat",'S')

[Walter Roberson]

Note this particular code assumes that the variable names are exactly the same for each file.

The output saved will be a non-scalar struct named S, with one field for each variable in the original files.

[Matt J]

Note this particular code assumes that the variable names are exactly the same for each file.

No, it doesn't assume that, but if there are non-overlapping variable names, it will result in spurious [] fields in the final struct array.

[Kishore]

Your code got worked.

Later, I ran the saved file of "JoinedData.mat".

It is showing the error: Unrecognized function or variable 'JoinedData'.

How to fix this error???

Can you answer or give me a lead to solve the problem???

Thank you

[Walter Roberson]

@Matt J

A = 3;
B = 5;
save data1.mat A
save data2.mat B

filenames = "data"+(1:2)+".mat";
for i=1:numel(filenames)
    S(i)=load(filenames(i));
end

Subscripted assignment between dissimilar structures.

load() returns a struct with one field for each variable name loaded. You cannot do subscripted assignment between structures with different field names -- or, for that matter between structures with different field names but different order.

[Walter Roberson]

@Kishore

You cannot "run" a .mat file. You can load() a .mat file.

[Kishore]

Oh, yeah!! sorry.

Actually, I commanded on,

Command window

load ('JoinedData.mat')

Error: Unrecognized function or variable 'JoinedData'.

Workspace

Name Value

S 1x10 struct with 70 fields

Now, what should I command to open the "JoinedData.mat" file????

[Kishore]

My .mat files matrix values are

Data1.mat is 22222 x 67

Data2.mat is 22222 x 67

JoinedData.mat matrix value should be like this, 44444 x 67.

[Matt J]

JoinedData.mat matrix value should be like this, 44444 x 67

It won't be. loading JoinedData.mat will give you the structure S, which you already have in the workspace.

[Walter Roberson]

Please show the output of

whos -file Data1.mat
whos -file Data2.mat

Answer 2

[Stephen23]

Fake data:

X = 1:3;
save data1.mat X
X = 4:6;
save data2.mat X
X = 7:9;
save data3.mat X
clear

Joining file data:

N = 3;
C = cell(1,N);
for k = 1:N
    F = sprintf('data%d.mat',k);
    C(k) = struct2cell(load(F));
end
Xnew = vertcat(C{:});
save('joineddata.mat','Xnew')

Checking:

S = load('joineddata.mat');
S.Xnew

ans = 3×3

1 2 3 4 5 6 7 8 9