Mailing-List: contact dev-help@avro.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@avro.apache.org
Received-SPF: pass (nike.apache.org: domain of bernd.fondermann@gmail.com
 designates 209.85.215.41 as permitted sender)
MIME-Version: 1.0
In-Reply-To: 
 <CALEq1Z_cKt_+8R-pH_KO5yUq+0yS=ABQ5Pbd0n1DDKMt6mf9BA@mail.gmail.com>
References: 
 <CAFk14guMLVy=FUVctYTkZckL=w9idyztPAzk3HNUFOKTXWB1pw@mail.gmail.com>
	<CALEq1Z85jx8PPim726Y=k_f-EWqdqFTewT_VUPcENyGoW1KooA@mail.gmail.com>
	<CAFk14gsPZJNYPv7mqEs3JtXipvLi4KTB60BxkEAgWsNc7dvGCA@mail.gmail.com>
	<CALEq1Z_cKt_+8R-pH_KO5yUq+0yS=ABQ5Pbd0n1DDKMt6mf9BA@mail.gmail.com>
Date: Wed, 8 May 2013 07:34:41 +0200
Message-ID: 
 <CAFk14gtSs3Zb+VpnDJt1B2JY2whKpeXDxR_y3h8cATg3sE2yKA@mail.gmail.com>
Subject: Re: Reading nested specific type
From: Bernd Fondermann <bernd.fondermann@gmail.com>
To: dev@avro.apache.org
Content-Type: multipart/alternative; boundary=001a11c333a067359a04dc2e4af8

--001a11c333a067359a04dc2e4af8
Content-Type: text/plain; charset=ISO-8859-1

On Tue, May 7, 2013 at 11:39 PM, Doug Cutting <cutting@apache.org> wrote:

> On Tue, May 7, 2013 at 12:46 PM, Bernd Fondermann
> <bernd.fondermann@gmail.com> wrote:
> > Thanks, the bug is fixed if I don't use a union.
>
> I found the problem with unions & updated the patch for AVRO-1320 to
> fix this too.  Please let me know if that solves things for you.
>

I'll give it a try.


>
> > However, if there are two enums in a union, could the deserialization
> > possibly work? How would Avro know which one was serialized?
>
> The position of the schema in the union is written preceding the
> value.  So a union of ["null","string"] is written as either a zero
> alone or a one followed by a string.  Similarly, enum symbols are
> written as their position in their schema.  For example, with the
> following two enum schemas:
>
> {"name":"X", "type", "enum", "symbols":["A","B","C"]}
> {"name":"Y", "type", "enum", "symbols":["A","C","E"]}
>
> Then the union schema ["X","Y"], an "A" from the first would be
> written as zero zero, while the "A" from the second would be one zero.
>
> Does that answer your question?
>
> I see.  - Thank you very much.

  Bernd

--001a11c333a067359a04dc2e4af8--