You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
ARROW-12714 added the str_to_title kernel and a basic mapping, but we should add a test. Also the stringr function takes a "locale" argument which is not handled here; we should either pass it to Arrow C++ if it supports it (which I doubt) or error if a value is provided in R.
This also applies to str_to_lower and str_to_upper kernels.
Eduardo Ponce / @edponce:
Arrow string compute functions do not support a locale setting and use the default POSIX locale which is "C".
Arrow provides UTF-8 variants of the string functions for localization standardization.
Currently, only strftime uses a locale option for formatting the resulting string.
Localization would also need to be considered for kernels that use/change text casing, compare strings, and format numbers.
Created ARROW-14126 to further investigate localization support for string functions.
Eduardo Ponce / @edponce:
If (for now) we are going to not accept locale option but still want to match stringr API, I can think of any of these approaches:
str_to_lower(x, locale = "en") -> only accept "en" locale and error if other
str_to_lower(x, locale = NULL) -> only accept NULL locale (ie., not supported) and error if another value is given
str_to_lower(x, locale) -> use missing() function to R to detect locale arg, if it is provided then error out
ARROW-12714 added the str_to_title kernel and a basic mapping, but we should add a test. Also the stringr function takes a "locale" argument which is not handled here; we should either pass it to Arrow C++ if it supports it (which I doubt) or error if a value is provided in R.
This also applies to str_to_lower and str_to_upper kernels.
Reporter: Neal Richardson / @nealrichardson
Assignee: Eduardo Ponce / @edponce
Related issues:
PRs and other links:
Note: This issue was originally created as ARROW-13853. Please see the migration documentation for further details.
The text was updated successfully, but these errors were encountered: