[ 
https://issues.apache.org/jira/browse/AVRO-2164?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16525022#comment-16525022
 ] 

Nandor Kollar commented on AVRO-2164:
-------------------------------------

As of schema evolution, Avro 
[spec|https://avro.apache.org/docs/1.8.2/spec.html#Decimal] has a rule for 
decimal evolution: the scale and precision should not change. Like Ryan 
mentioned this can result in wrong interpretation of data, therefore in my 
opinion the code should enforce this rule (see AVRO-2078). 

However spec doesn't have any evolution rule for other types. What happens if 
the millisecond precision timestamp is promoted to microsecond precision? 
Should Avro care about it? What happens to the other way? What happens if 
reader tries to read a time field as date?

> Make Decimal a first class type.
> --------------------------------
>
>                 Key: AVRO-2164
>                 URL: https://issues.apache.org/jira/browse/AVRO-2164
>             Project: Avro
>          Issue Type: Improvement
>          Components: logical types
>    Affects Versions: 1.8.2
>            Reporter: Andy Coates
>            Priority: Major
>
> I'd be interested to hear the communities thoughts on making decimal a first 
> class type. 
> The current logical type encodes a decimal into a _bytes_ or _fixed_. This 
> encoding does not include any information about the scale, i.e. this encoding 
> is lossy. 
> There are open issues around the compatibility / evolvability of schemas 
> containing decimal logical types, (e.g. AVRO-2078 & AVRO-1721), that mean 
> reading data that was previously written with a different scale will result 
> in data corruption.
> If these issues were fixed, with suitable compatibility checks put in place, 
> this would then make it impossible to evolve an Avro schema where the scale 
> needs to be changed. This inability to evolve the scale is very restrictive, 
> and can result in high overhead for organizations that _need_ to change the 
> scale, i.e. they may potentially need to copy their entire data set, 
> deserializing with the old scale and re-serializing with the new.
> If _decimal_ were promoted to a first class type, this would allow the scale 
> to be captured in the serialized form, allow for schema evolution support.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to