[
https://issues.apache.org/jira/browse/DRILL-7370?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17004139#comment-17004139
]
Arina Ielchiieva commented on DRILL-7370:
-----------------------------------------
Charles, dynamic UDFs is rather complex project that consumes jar with UDFs and
then distributes them locally to each Drillbit, saving UDFs signatures and
metadata in the Zookeeper. Frankly saying I don’t think such approach would
suit your needs plus will bring additional load to the Zookeeper. For your
needs, you might consider using Metastore, developing a module that would be
responsible for models storage and retrieval but I believe this would require
some design first. The simplest solution is for you instead of indicating jar
in function call, include full path to jar. If Drill is run locally, this can
be path on local file system, if you have several drillbits, than path should
be on distributed file system: {{predict(‘file:///path/to/jar/my_jar.jar’,
....)}}. Thus user won’t have to deal with jar registration, distribution,
storage etc.
> Add Generic Predict UDF for H20 ML Models
> -----------------------------------------
>
> Key: DRILL-7370
> URL: https://issues.apache.org/jira/browse/DRILL-7370
> Project: Apache Drill
> Issue Type: Improvement
> Components: Functions - Drill
> Affects Versions: 1.17.0
> Reporter: Charles Givre
> Assignee: Charles Givre
> Priority: Major
>
> h20.ai enables a user to export a trained machine learning model as either a
> POJO or MOJO. The proposed UDF will implement the `predict()` function and
> enable ML predictions to be included in Drill queries.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
