Native pipe does not work with rowwise

[Yann-C-INN]

Using the dplyr::rowwise i noticed something strange.
The native R pipe (|>) Doesn't work, where's the magrittr::%>% works just fine.

The following code works:

mtcars |> rowwise() %>% mutate(sum = sum(.))

But this one doesnt:

mtcars |> rowwise() |> mutate(sum = sum(.))

[bhattmaulik]

I guess the problem is not the pipe, but the dot inside sum(.). The base R pipe needs a different syntax to work.

[Yann-C-INN]

I guess the problem is not the pipe, but the dot inside sum(.). The base R pipe needs a different syntax to work.

It doesnt make sensethat these pipes will not work the same... I find this a bugrather than a feature.

For my own purpose I already solved it (My use case was a bit more complex obviously), but it doesnt resonate with me that they would behave differently.

[philibe]

It doesnt make sensethat these pipes will not work the same... I find this a bugrather than a feature.

I think you're wrong :)

They are similar but intrinsically a little different.

Read that please:

• https://www.tidyverse.org/blog/2023/04/base-vs-magrittr-pipe/
• https://towardsdatascience.com/understanding-the-native-r-pipe-98dea6d8b61b
• https://stackoverflow.com/questions/67633022/what-are-the-differences-between-rs-native-pipe-and-the-magrittr-pipe/72086492#72086492

edit: the Stackoverflow answer added

[philibe]

Here is a workaround for an identical result to your example mtcars |> rowwise() %>% mutate(sum = sum(.))

• mtcars |> {\(.) mutate(., sum= sum(.)) }() inspired from https://stackoverflow.com/questions/75961573/use-base-r-pipe-in-conditional-dplyr-chain/75961590#75961590

[Yann-C-INN]

Here is workaround for result identical to your example mtcars |> rowwise() %>% mutate(sum = sum(.))

• mtcars |> {\(.) mutate(., sum= sum(.)) }() from https://stackoverflow.com/questions/75961573/use-base-r-pipe-in-conditional-dplyr-chain/75961590#75961590

the workaround is to use the more robust pipe if you ask me ;) thanks though. I am not convinced that this is something worth keeping in the future (the differences) but sure. I'll adjust my defaults

[DavisVaughan]

The base R pipe is deliberately different from the magrittr pipe. The differences are outlined in https://www.tidyverse.org/blog/2023/04/base-vs-magrittr-pipe/#-vs as @philibe mentioned, and there is nothing dplyr can do to support this differently