Can MATLAB read 'R' data files (*.Rda) directly?

If so, what is the appropriate function?

Answer 1

[Peter]

If you have Python installed on your system, you can use use the Matlab python call functionality to read in such R files.

Firstly, ensure that you have installed the Python package pyreadr onto your system; this package reads/writes R RData and Rds files into/from pandas data frames, and does not require R to have been installed, not any other external dependencies installed.

Then you can load in matlab via the command:

dict = py.pyreadr.read_r(<path to rda file>);

As it stands the variable "dict" will be a Python OrderedDict, so to use data it contains you need to create a matlab array as follows:

result=double(dict{<key>}.values);

where <key> is the string defining which key of the pandas frame dict you wish to extract.

If in doubt, check which keys the imported panda frame has by running the command:

dict.keys()

Once you have found the key of interest, you may further check the complete properties of the data stored under this key by running the command:

dict{<key>}

Answer 2

[Pierre Benoit]

Apparently no, see Supported File Formats for Import and Export.

The only other topic I found about this was this.

[Richard]

Thanks, Pierre. I ended up reading them in to 'R', then saving as a CSV which imports easily into MATLAB. Since I am new to 'R', it took me some time to generate the appropriate code. Here is an example:

> load(“kidneydata.Rda”) % from Brad Efron's web site

> dim(kidneydata) % returns 157 x 2

> A <- t(kidneydata) % transpose array

> dim(A) % returns 2 x 157

> write(A,"kidneydata.csv",ncolumns = 2,sep = ",") % save as CSV file

> write(A,"kidneydata.csv",ncolumns = 2,sep = "\t") % to save as TSV file

[Andrew]

Now in 2020 with R studio 1.1.456, the 'write' function croaked with a mysterious "Error in cat...argument 1 (type 'list') cannot be handled by cat" error, but:

write.csv(variable,"filename");

worked just fine.