Davis Vaughan comments

Results 682 comments of


                                            Davis Vaughan

trafficstars

Error in 'vec_rbind()'

Could you please turn this into a self-contained [reprex](http://reprex.tidyverse.org/) (short for minimal **repr**oducible **ex**ample)? It will help us help you if we can be sure we're all working with/looking at...

Can this joint-column `pivot_longer()` situation be simplified?

No, but ill probably take an in depth look at this once we do the next minor release of tidyr, probably in a few months

Any possibility of being less strict with sequence methods?

The main issue with this comes with the `iso_year_week_day -> naive_time` conversion. This is only _always_ possible for this class at "year" precision (for `year_month_day` that conversion is possible up...

Calendar: Solar Hijri / Persian

Oh it lives here now https://github.com/r-lib/tzdb/blob/main/inst/include/date/solar_hijri.h

Calendar: Solar Hijri / Persian

I'm just saying that the C header file now lives in that other package. I haven't added R bindings to it here yet. It isn't very high priority for me...

Feature request: custom `direction` in `new_metric()`

The biggest issue with a completely custom `direction` is that tune somehow has to know how to use it. In particular, `tune_bayes()` does different things depending on whether you are...

[Feature request or question] Provide option for metric analysis by class in a multi-class setting

I think this is the same as https://github.com/tidymodels/yardstick/issues/4 i.e. it sounds like you want one-vs-all output (with regard to factor levels). Like: ```r # truth "a" "b" "c" "c" "a"...

[Feature request or question] Provide option for metric analysis by class in a multi-class setting

You cannot currently do it. I am not sure if we should expose these utilities or not, but it has come up a few times so it might be worth...

Recompile if `.o` and `.so` are from a different arch or R version

See also https://github.com/r-lib/pkgload/issues/226

`date_parse()` (and friends) should get `failure = c("warn", "error", "NA")`

In some cases, we want to "try to parse" and detect which rows would fail to parse to identify them as "problems" (and maybe you'd throw those rows out entirely...