brendanjodowd / sas-r

SAS to R cheatsheet

Hi there 👋

I am a Senior Statistician at the Central Statistics Office (CSO). I use GitHub for a couple of work-related and hobby projects. My main languages are R and SAS.

Some projects

• My GitHub Pages site where I have guides for making maps and implementing statistical disclosure controls.
• 🎓 Higher Education Outcomes - a longitudinal study of outcomes for higher education graduates carried out using admin data.
• 🌐 Ireland maps - a collection of compatible maps produced using shapefiles from OSi and OSNI and the rmapshaper package.
• 💑 SAS-R Cheatsheet - a guide to learning R for SAS users, or vice versa. Now hosted on the RStudio website.
• 📈 R Package for importing data from CSO Statbank - This has been superseded by an official CSO package.
• 👨‍👩‍👧‍👦 LEA Population Map - A Shiny web app to calculate population in selected LEAs
• 🍫 The Axis of Fudge - Another Shiny project to interactively examine and compare brownie recipes
• 🔧 SAS Macros - a ton of macros for SAS that I developed/collected over the years

You can find me on LinkedIn here.

sas-r's Issues

Save/Load examples

I love your cheat sheet for introducing SAS users to R, but I'd make one suggestion:

In the File operations suggestion, I believe you should replace the save()/load() examples with saveRDS() and readRDS()

The *RDS functions are functionally equivalent to setting SAS data into/from a library--they store a single object and they let you easily define a name on load.

Generally using save() is a bad habit, especially when sharing data with others and only saving a single dataframe. It can be useful in certain circumstances (e.g. saving many pieces of an environment for resumption later) but I'd say that's beyond the scope of a new-user cheat sheet.

Load() can be particularly dangerous since it will overwrite existing objects without warning. It can be potentially catastrophic to perform a plain load() on a file created by another team member. Especially for SAS users who may be new to concepts like environments, variable scopes, and the existence of variables/objects beyond dataframes. Much better to use loadRDS() where you can pick and choose which objects you actually want to load and assign them non-conflicting names.