You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@beam.apache.org by "Luke Cwik (JIRA)" <ji...@apache.org> on 2016/12/28 21:54:58 UTC

[jira] [Created] (BEAM-1226) Use well known coder types in Python for KV, Iterable, WindowedValue, GlobalWindow, and LengthPrefix

Luke Cwik created BEAM-1226:
-------------------------------

             Summary: Use well known coder types in Python for KV, Iterable, WindowedValue, GlobalWindow, and LengthPrefix
                 Key: BEAM-1226
                 URL: https://issues.apache.org/jira/browse/BEAM-1226
             Project: Beam
          Issue Type: Improvement
          Components: beam-model-fn-api
            Reporter: Luke Cwik
            Assignee: Luke Cwik
            Priority: Minor


Add support for using kind for well known coder types in Apache Beam for Python

The following specific cloud object representations for the following types should exist:
kind:pair (TupleCoder with two components)
kind:stream (IterableCoder with a single component)
kind:global_window (GlobalWindowCoder)
kind:length_prefix (A new type of coder which always encodes the length of the value type as a prefix, has a single component)
kind:windowed_value (A wrapper coder with two components (value coder and window coder))

These changes are towards having a common binary representation for certain well known coders across multiple languages.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)