Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Avro >> mail # user >> Best Practices - c#/java/ruby, decimals and dates

Copy link to this message
Best Practices - c#/java/ruby, decimals and dates

I have a legacy windows c# implementation that needs to be able serialize a
large object structure for deserialization in java and ruby.  This contains
business data and has many fields that are either decimal or date/datetime

To facilitate this, I am creating an Avro json schema that can be used by
c#, java, and ruby implementations.  My question is about the best way to
handle decimal and date datatypes.  Yes, I know they could be converted to
strings for serialization, but that would require explicit conversion to and
from the native decimal/bigdecimal types and in general force the Avro
emitted classes to be either wrapped or duplicated in classes which expose
them as the desired decimal or date types.

Since Avro does not natively support these types, what is considered best
practice?  I'd certainly like to avoid the overhead/complexity of wrapping
an entire class structure.  I see posts about java annotations for date
types that look like they will be in 1.7.5, but I don't see how this will
help in my situation.

Thanks in advance,

View this message in context: http://apache-avro.679487.n3.nabble.com/Best-Practices-c-java-ruby-decimals-and-dates-tp4027985.html
Sent from the Avro - Users mailing list archive at Nabble.com.