In tidyr 1.2.0 we updated pivot_longer()'s names_transform, values_transform, names_ptypes and values_ptypes arguments to allow a single function or ptype which will be applied to all columns, rather than requiring a named list. See https://github.com/tidyverse/tidyr/blob/main/NEWS.md#pivoting
This required us to change the default argument from list() to NULL. In particular, for names/values_ptypes, this was important because list() should now be interpreted as "use a list ptype for all columns".
For backwards compatibility with tidyseurat and a few other packages that extend pivot_longer(), we have temporarily forced list() to be interpreted as NULL, but we would like to remove this behavior. See https://github.com/tidyverse/tidyr/issues/1296.
This PR is a step towards that by updating the signature used in tidyseurat to use NULL rather than list(). We would appreciate if you could merge this in and get it on CRAN at your earliest convenience!
In tidyr 1.2.0 we updated
pivot_longer()
'snames_transform
,values_transform
,names_ptypes
andvalues_ptypes
arguments to allow a single function or ptype which will be applied to all columns, rather than requiring a named list. See https://github.com/tidyverse/tidyr/blob/main/NEWS.md#pivotingThis required us to change the default argument from
list()
toNULL
. In particular, fornames/values_ptypes
, this was important becauselist()
should now be interpreted as "use a list ptype for all columns".For backwards compatibility with tidyseurat and a few other packages that extend
pivot_longer()
, we have temporarily forcedlist()
to be interpreted asNULL
, but we would like to remove this behavior. See https://github.com/tidyverse/tidyr/issues/1296.This PR is a step towards that by updating the signature used in tidyseurat to use
NULL
rather thanlist()
. We would appreciate if you could merge this in and get it on CRAN at your earliest convenience!