Best way to simply store data for statistical analysis in R I have been using text files to store my data for R without any problem for some time now. But for a recent project the files' sizes are getting too big for raw text files to handle. What's the best simple alternative?
 A: Take a look at SQLite3 databases. Each database is a file, so it doesn't require setting up a database server.
To create a database:
$ sqlite3 my_db.db3
> CREATE TABLE my_table ( col1 TEXT );

For use with R, https://gist.github.com/lynaghk/1062939
A: There are a number of generic options.  


*

*You could compress the text.  

*You could go binary on the text, not write in ascii


Great compression is data dependent.
My guess (and you didn't specify so I must guess) is that you are looking to store spreadsheet-like data in something other than csv (comma-delimited).
One of my favorite formats (I love MatLab) is hdf.  
Here is R-related information about HDF:


*

*accessed through packages hdf5, h5r, Bioconductor's rhdf5, RNetCDF, ncdf and ncdf4 

*http://www.hdfgroup.org/HDF5/whatishdf5.html
It is a high-density supercomputing data storage format.  It can be very fast and efficient.  It is also (unsurprisingly) denser than zipped text.
A: Standard file reading functions in R will now automatically easily read gzipped files. So, just run simple gzip compression on your data and read as always, as if it was plain text.
read.table('myfile.gz')

A: The standard R approach is to use save and load.  If you run save on your data frame after importing and annotating it, you can specify compress=TRUE and you'll be amazed at the compression and the fast load time.  This works especially well if the object size is less than about 400MB.  Otherwise, check out some of the suggestions above, or the powerful ff package in R.
The Hmisc package has little wrappers Save and Load to make the above even more painless:
mydata <- csv.get(...)  # Hmisc package, has several options
Save(mydata)   # writes mydata.rda to current working directory
....
Load(mydata)   # reads mydata.rda and creates mydata data frame

