[
https://issues.apache.org/jira/browse/ARROW-7844?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17035820#comment-17035820
]
Neal Richardson edited comment on ARROW-7844 at 2/13/20 12:22 AM:
------------------------------------------------------------------
Another failure, locally:
{code}
── 3. Failure: Lists are preserved when writing/reading from Parquet (@test-parq
`object` not equivalent to `expected`.
Component “int”: Component 1: target is numeric, current is logical
Component “int”: Component 2: target is numeric, current is logical
{code}
To be clear, this is looking at the first and second "rows" of the {{int}}
column in the data.frame.
And the next time:
{code}
── 1. Failure: Lists are preserved when writing/reading from Parquet (@test-parq
`object` not equivalent to `expected`.
Component “bool”: Component 2: target is logical, current is numeric
Component “int”: Component 1: target is numeric, current is logical
{code}
was (Author: npr):
Another failure, locally:
{code}
── 3. Failure: Lists are preserved when writing/reading from Parquet (@test-parq
`object` not equivalent to `expected`.
Component “int”: Component 1: target is numeric, current is logical
Component “int”: Component 2: target is numeric, current is logical
{code}
To be clear, this is looking at the first and second "rows" of the {{int}}
column in the data.frame.
> [R] Parquet list column test is flaky
> -------------------------------------
>
> Key: ARROW-7844
> URL: https://issues.apache.org/jira/browse/ARROW-7844
> Project: Apache Arrow
> Issue Type: Bug
> Components: R
> Reporter: Neal Richardson
> Assignee: Francois Saint-Jacques
> Priority: Major
>
> See
> [https://travis-ci.org/ursa-labs/arrow-r-nightly/jobs/649649349#L373-L375]
> for an example on public CI. I was seeing this locally this week but figured
> I'd screwed up my env somehow.
> {code}
> ── 1. Failure: Lists are preserved when writing/reading from Parquet
> (@test-parq
> `object` not equivalent to `expected`.
> Component "num": Component 1: target is numeric, current is character
> {code}
> It's not always the same column in the data.frame that is affected. Also
> strange that it's only one column. You'd think that if it were transposing
> the order somehow, you'd get two that were swapped.
> The test itself is straightforward
> (https://github.com/apache/arrow/blob/master/r/tests/testthat/test-parquet.R#L124-L137)
> so this is somewhat troubling.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
