RAthena 1.3.0 has arrived
RAthena is a R package that interfaces into Amazon Athena. However, it doesn’t use the standard
JDBC drivers like
RAthena utilises Python’s SDK (software development kit) into Amazon,
Boto3. It does this by using the
reticulate package that provides an interface into Python. What this means is that
RAthena doesn’t require any driver installation or setup. That can be particularly difficult when you are considering setting up the ODBC drivers and you are not familiar with how ODBC works on your current operating system. If you wish to use ODBC, RStudio has provided a good user guide Setting up ODBC Drivers to help set up ODBC drivers on your system. However if you do not wish to go down that route
RAthena might be a good option for you.
New Features in
Anyway, getting back to
RAthena and what does the new update provide. One of the key changes in
RAthena is the method of transferring data to and from AWS Athena.
RAthena now utilising
data.table for this process. The reason for this change is the raw speed
data.table. When transferring data to and from AWS Athena the last thing you want is a bottle neck in R just preparing the data before it even transfers it to AWS Athena. This bottle neck can easily be 50 - 100x longer without the use of
The next change is
bigint, and how it is converted from AWS Athena to R. In the past
RAthena would just convert
bigint when writing to AWS Athena, however it would then convert
bigint back into R as a normal
integer. Which means it is constrained to 32-bit integers. This has now been fixed. When reading
bigint from AWS Athena
RAthena will now convert it into
RAthena now provides a faster method in reading and writing data from AWS Athena (thanks
data.table). With the correct handling of AWS Athena
bigint. So please give
RAthena a try and let me know what you think of the package. Suggestions/Bugs/Enhancements are always welcome and they will help the package to improve: https://github.com/DyfanJones/RAthena/issues.
Just in case you are not aware
Rathena is available on the CRAN and GitHub.
GitHub development version: