remove files greater than 100MB from github repo before committing

I accidentally commit files greater than 100 MB (which is against GitHub’s policy - only files less than 100MB). Hence, you should always either

  1. Ignore the files by including their names in the .gitignore files. You can either write the name of the file or use * (e.g., *.rds to ignore any files with rds ending).
  2. Use Git Large File Storage to store the files. This step is covered in this post

Back to our problems, how can we delete the wrong commits?

Luckily we can use bfg.jar, which could be downloaded from here. I usually move the file to my GitHub directory that I want to use. But you can also leave it wherever you want.

To delete the wrong commits, you can either go to “Terminal”, or I usually just to the “Terminal” tab in RStudio. Then, you can type

java -jar bfg.jar --strip-blobs-bigger-than 100M <your_repo>

For example, I use

java -jar bfg-1.14.0.jar --strip-blobs-bigger-than 100M C:/Users/tn9k4/GitHub/finance

where I have to change bfg.jar because my file has its version in its name (or you can also change your bfg.jar to match with your command) and my git repo is C:/Users/tn9k4/GitHub/finance

To find your git repo, you can also type in the “Terminal” tab

git rev-parse --show-toplevel

Please don’t include .git in the repo location.

Mike Nguyen, PhD
Mike Nguyen, PhD
Visiting Professor of Data Sciences and Operations

My research interests include marketing, and social science.

Next
Previous

Related