I'd run into a similar issue, where I used a record name with a "-" in it when 
creating files with Avro 1.4.

So I've got files that are valid/readable with 1.4, but seem to be unusable 
with 1.5.

-- Ken

On Jun 10, 2011, at 12:32pm, Matt Pouttu-Clarke wrote:

> OK, thanks.  This is a change from Avro 1.4 which will make older Avro files
> unreadable by 1.5 until the field names are changed.
> 
> Can we get away with "tweeking" the field name in the JSON meta data portion
> of the file, or do we have to re-write the whole file again with the new
> meta data?
> 
> On 6/10/11 12:23 PM, "Douglas Creager" <[email protected]> wrote:
> 
>>> I am using Avro for processing/encoding billions of rows using Hadoop.  When
>>> I attempted to upgrade to Avro 1.5 I got an error about illegal field names.
>>> I notice this case happens when I use a O-’ character or a O.’ in a field
>>> name.  
>>> 
>>> Are there any other restrictions on fields names I need to know when
>>> upgrading?
>> 
>> You can find the full details in the Avro spec [1].  It's basically just like
>> the identifier restrictions in most programming languages.  Names can only
>> contain letters, numbers, and underscores, and can't start with a number.
>> Dots are used to created namespaces, just like fully-qualified Java
>> identifiers.
>> 
>> [1] http://avro.apache.org/docs/current/spec.html
>> 
>> cheers
>> –doug
> 
> 
> iCrossing Privileged and Confidential Information
> This email message is for the sole use of the intended recipient(s) and may 
> contain confidential and privileged information of iCrossing. Any 
> unauthorized review, use, disclosure or distribution is prohibited. If you 
> are not the intended recipient, please contact the sender by reply email and 
> destroy all copies of the original message.
> 
> 

--------------------------
Ken Krugler
+1 530-210-6378
http://bixolabs.com
custom data mining solutions






Reply via email to