[
https://issues.apache.org/jira/browse/MADLIB-1024?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15867004#comment-15867004
]
Rahul Iyer commented on MADLIB-1024:
------------------------------------
[~mausch] Thanks for the extension file. My take on the error is that it's
caused because {{svec_util.sql}} is placed into that file before {{svec.sql}}.
I don't yet have an answer on why that is the case. A topological sort is
performed based on a dependency graph between the modules - this sort should
put svec before svec_util.
Please share all environment information that you're running the docker in. OS,
Python version, etc? Maybe there is a bug in the sort code that is being
triggered in certain situations.
> CREATE EXTENSION madlib fails
> -----------------------------
>
> Key: MADLIB-1024
> URL: https://issues.apache.org/jira/browse/MADLIB-1024
> Project: Apache MADlib
> Issue Type: Bug
> Reporter: Martin Jensen
> Assignee: Orhan Kislal
> Fix For: v1.10
>
> Attachments: Dockerfile
>
>
> When installing madlib with PGXN on PG 9.5.4, Ubuntu 16.04 I get the
> following error when trying to create the extension:
> ERROR: aggregate cannot accept shell type public.svec
> ********** Error **********
> ERROR: aggregate cannot accept shell type public.svec
> SQL state: 42P13
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
