[
https://issues.apache.org/jira/browse/ARROW-1254?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Wes McKinney resolved ARROW-1254.
---------------------------------
Resolution: Not A Problem
Because of the way conda-forge works (takes a long time to update multiple
interdependent packages), this is usually going to happen around release time.
The way to avoid it is to pin the package version rather than installing
"head". I think when we do "pre" builds on conda-forge we should pin the pre
build so that when we update to final it doesn't break things if pyarrow (last
package in the dependency chain) hasn't been updated
> failing installation on osx / linux
> -----------------------------------
>
> Key: ARROW-1254
> URL: https://issues.apache.org/jira/browse/ARROW-1254
> Project: Apache Arrow
> Issue Type: Bug
> Affects Versions: 0.5.0
> Reporter: Jeff Reback
> Assignee: Uwe L. Korn
>
> This might be pre-mature as it looks like arrow-cpp is the released version,
> but pyarrow is still a pre-build (maybe still building)?
> xref https://github.com/pandas-dev/pandas/issues/17064
> This install is on osx, but seeing issue on travis (linux)
> ```
> The following packages will be downloaded:
> package | build
> ---------------------------|-----------------
> arrow-cpp-0.5.0 | np113py36_0 1.8 MB conda-forge
> python-dateutil-2.6.0 | py36_0 236 KB defaults
> pyarrow-0.5.0.pre | np113py36_2 511 KB conda-forge
> ------------------------------------------------------------
> Total: 2.6 MB
> The following NEW packages will be INSTALLED:
> python-dateutil: 2.6.0-py36_0 defaults
> The following packages will be UPDATED:
> arrow-cpp: 0.4.1-np113py36_2 conda-forge --> 0.5.0-np113py36_0
> conda-forge
> pyarrow: 0.4.1-np113py36_0 conda-forge --> 0.5.0.pre-np113py36_2
> conda-forge
> The following packages will be DOWNGRADED due to dependency conflicts:
> jemalloc: 5.0.0-0 conda-forge --> 4.4.0-0
> conda-forge
> Proceed ([y]/n)? y
> ```
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
