Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Wrong results in "summarise" when characters are used as variable identifiers #567
The following code
doesn't rise any errors or warning but returns (randomly) wrong results (but approximately the same magnitude) between 16 and 24, whereas the correct use of the variable.
Indicated by the stackoverflow question "Randomness in dplyr" a similar behavior occurs on larger datasets too.
The problem is there somewhere:https://github.com/hadley/dplyr/blob/424b304ae235d9ff33b6fbcd7860e8f94ec7e9f2/inst/include/dplyr/Result/Count_Distinct.h
@hadley : Should I make
By the same token, would we allow expressions inside of
This is because the underlying class
I'm not trying to push towards forbidding things, but it would be easier to name whatever expression that would go inside