R doesn’t need to throttle AWS Athena anymore

[This article was first published on Dyfan Jones Brain Dump HQ, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

RBloggers|RBloggers-feedburner

I am happy to announce that RAthena-1.9.0 and noctua-1.7.0 have been released onto the cran. They both bring two key features:

  • More stability when working with AWS Athena, focusing on AWS Rate Exceeded throttling errors
  • New helper function to convert AWS S3 backend files to save cost

NOTE: RAthena and noctua features correspond to each other, as a result I will refer to them interchangeability.

Stability

Throttling AWS

One of the main problems when working with AWS API is stumbling into Rate Exceeded throttling error. With the latest update to the packages, the connection between AWS Athena and R has been made more robust through retry functionality. This allows R to automatically retry its request using an exponential backoff with jitter (Best practices for API packages)

  1. R sends a call to AWS Athena, let’s say dbListTables(con). However AWS is over run with requests, and returns an error back to R saying it is overwhelmed (this is a rate exceeded throttling error). As RAthena and noctua retry noisely, the error will be printed to the console letting you know AWS is busy ({expection message} + "Request failed. Retrying in " + {wait time} + " seconds...").
  2. R will then wait for a given time (please see error format above) and retry the request again. AWS replies it is still busy and can’t do the request.
  3. This time R will back off for a long period of time, this gives AWS some breathing room. Now when R sends the request over to AWS, AWS is able to complete the call and return out desired results.

This feature is a great step in the right direction for making R and AWS Athena work together seamlessly. For anyone who wishes to create their own retry method both packages have enabled this through their ..._options() function. For more information please refer to link.

Save the pennies

Converting AWS S3 files

AWS Athena costs by the amount of data it scans. This makes it very important to have your AWS S3 backend files in the suitable format to reduce the cost of using AWS Athena. This is where the next key feature comes in. This feature basically creates a simple wrapper to allow you to convert AWS S3 files into a more suitable format.

library(DBI)
library(RAthena)

con <- dbConnect(athena())

# Upload iris data.frame to AWS Athena as a delimited file
dbWriteTable(con, "iris_delim", iris)

# Convert to parquet using AWS Athena
dbConvertTable(con,
               obj = "iris_delim",
               name = "iris_parquet",
               file.type = "parquet")

In this example simply uploaded iris data.frame to AWS Athena in a default delimited file format (please see link for more information around how to upload data to AWS Athena). Then it is converted into parquet file format using AWS Athena. This wrapper isn’t limited to converting just AWS Athena tables, it can also convert SQL DML queries. Please refer to dbConvertTable for more documentation or to the dbConvertTable vignette link.

Finally for more informations around best practises with AWS Athena please look at Top 10 Perfromace Tuning Tips for Amazon Athena

Sum Up

These two new features bring R and AWS Athena that little bit closer together. As always if you have any new features or identify any bugs please feel free to raise a pull request or ticket on the corresponding package github pages (RAthena and nocuta)

To leave a comment for the author, please follow the link and comment on their blog: Dyfan Jones Brain Dump HQ.

R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

Never miss an update!
Subscribe to R-bloggers to receive
e-mails with the latest R posts.
(You will not see this message again.)

Click here to close (This popup will not appear again)