gora-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Talat UYARER (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (GORA-279) Hazelcast module
Date Sat, 29 Mar 2014 17:26:15 GMT

    [ https://issues.apache.org/jira/browse/GORA-279?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13954350#comment-13954350

Talat UYARER commented on GORA-279:

Hi all,

I agree with [~alparslan.avci]'s avro opinion. We use avro's compiler and objects. If we use
avro's serializers, we can easily develop most of thing. But I think our first priority should
be readable and accessible data without Gora's or Avro's serializers. Default serializers
of avro store union's using type at first byte of value. So Union types become broken after
the serialize if we use standard avro's seriliazers. All nullable fields are union fields.
IMHO This is a big problem. 

I have a solution for this problem. My solution is taking snapshot of object when persist
data to store. The snapshot store types of every field for the object, specially for union
fields. We write this snapshot as a bytebuffer out of schema field. And then when creating
object from the store we can use the snapshot. 

If we will be glad this solution, Maybe we develop for avro serilaizers.

Wdyt guys ?

> Hazelcast module
> ----------------
>                 Key: GORA-279
>                 URL: https://issues.apache.org/jira/browse/GORA-279
>             Project: Apache Gora
>          Issue Type: Wish
>            Reporter: Julien Nioche
>              Labels: gsoc2014
>             Fix For: 0.5
>         Attachments: gsoc_proposal_draft.odt
> HazelCast seems to be a popular in-memory data grid [http://www.hazelcast.com/]. The
community edition is under the ASL v2.

This message was sent by Atlassian JIRA

View raw message