[
https://issues.apache.org/jira/browse/ARROW-12322?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ian Cook updated ARROW-12322:
-----------------------------
Description:
There are more than two dozen data type functions in the arrow package, and
they are named very generically, so they represent a large surface area for
potential masking problems, which are likely to occur in user environments, not
in our CI. If these masking errors do occur, they will probably give
frustratingly unhelpful error messages. This happened to me with
{{rlang::string()}}. The error was:
{quote}Error in is_integerish\(x\) : argument "x" is missing, with no default
{quote}
This can be worked around with some non-standard eval magic.
I implemented a working version of this in
[https://github.com/apache/arrow/pull/9952] but we removed it before merging
that PR because there were questions about whether there was a better way to
implement it.
was:
There are more than two dozen data type functions in the arrow package, and
they are named very generically, so they represent a large surface area for
potential masking problems, which are likely to occur in user environments, not
in our CI. If these masking errors do occur, they will probably give
frustratingly unhelpful error messages. This happened to me with
{{rlang::string()}}. The error was:
{quote}Error in is_integerish(x) : argument "x" is missing, with no default
{quote}
This can be worked around with some non-standard eval magic.
I implemented a working version of this in
[https://github.com/apache/arrow/pull/9952] but we removed it before merging
that PR because there were questions about whether there was a better way to
implement it.
> [R] Work around masking of data type functions
> ----------------------------------------------
>
> Key: ARROW-12322
> URL: https://issues.apache.org/jira/browse/ARROW-12322
> Project: Apache Arrow
> Issue Type: Improvement
> Components: R
> Reporter: Ian Cook
> Assignee: Ian Cook
> Priority: Major
>
> There are more than two dozen data type functions in the arrow package, and
> they are named very generically, so they represent a large surface area for
> potential masking problems, which are likely to occur in user environments,
> not in our CI. If these masking errors do occur, they will probably give
> frustratingly unhelpful error messages. This happened to me with
> {{rlang::string()}}. The error was:
> {quote}Error in is_integerish\(x\) : argument "x" is missing, with no default
> {quote}
> This can be worked around with some non-standard eval magic.
> I implemented a working version of this in
> [https://github.com/apache/arrow/pull/9952] but we removed it before merging
> that PR because there were questions about whether there was a better way to
> implement it.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
