[
https://issues.apache.org/jira/browse/ARROW-12761?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17343539#comment-17343539
]
Neal Richardson commented on ARROW-12761:
-----------------------------------------
The error is actually not in write_to_raw, I don't think. It's probably here:
https://github.com/apache/arrow/blob/master/r/R/ipc_stream.R#L48
I think you fix it by adding validation in L42 that x is ArrowTabular. You
might want to check for this validation logic gap in the other write_*
functions too.
> [R] Better error handling for write_to_raw
> ------------------------------------------
>
> Key: ARROW-12761
> URL: https://issues.apache.org/jira/browse/ARROW-12761
> Project: Apache Arrow
> Issue Type: Improvement
> Components: R
> Reporter: Nic Crane
> Priority: Minor
>
> When I run this:
> {code:java}
> mt_array <- Array$create(mtcars)
> write_to_raw(mt_array){code}
> I get this error:
> {code:java}
> Error: schema must be a "Schema" {code}
> We should add a manual check of the type of the input that returns a more
> informative error message if the object is not a data.frame, RecordBatch or
> Table
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
