Attributes lost on `dplyr::mutate` and `dplyr::filter`

[ashiklom]

These are common operations that should really preserve attributes. However, because of the way they are implemented in dplyr (with an underlying call to C code), all attributes are lost following these operations.

I think a similar pattern to the one you currently have -- extract the attributes, call NextMethod, and re-assign the attributes -- should work fine, so it should be straightforward to implement.

[ctbrown]

​Thanks for submitting the issues and using sticky. I have noticed that a number of dplyr/tidyverse operations do not preserve attributes. I am trying to find a more general method for preserving attributes for these operations, but think that you are correct, that the NextMethod method is probably the most viable solution.

…

On Sun, Mar 25, 2018, 10:08 AM Alexey Shiklomanov ***@***.***> wrote: These are common operations that should really preserve attributes. However, because of the way they are implemented in dplyr (with an underlying call to C code), all attributes are lost following these operations. I think a similar pattern to the one you currently have -- extract the attributes, call NextMethod, and re-assign the attributes -- should work fine, so it should be straightforward to implement. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#7>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AC5xawNvE8BWiaMSCQgEv_S-FDOTIo8Gks5th877gaJpZM4S6N8e> .

[maxheld83]

I am trying to find a more general method for preserving
attributes for these operations

see:

• adv-r (discusses dplyr::mutate() in particular)
• sloop::reconstruct(), also in sticky::sticky() + sloop::reconstruct() = bliss? #10