[jira] [Created] (BEAM-700) Update FileBasedSource to use sampling when estimating size of file patterns

["Chamikara Jayalath (JIRA)" <ji...@apache.org>]

Chamikara Jayalath created BEAM-700:
---------------------------------------

             Summary: Update FileBasedSource to use sampling when estimating size of file patterns
                 Key: BEAM-700
                 URL: https://issues.apache.org/jira/browse/BEAM-700
             Project: Beam
          Issue Type: Bug
          Components: sdk-py
            Reporter: Chamikara Jayalath
            Assignee: Chamikara Jayalath
             Fix For: Not applicable


Currently when estimating size of file patterns we estimate size of each file and add the sizes up [1]. This can be very inefficient for file patterns that expand into a large number of files. We should update FIleBasedSource to use sampling similar to Java SDK [2].

[1]https://github.com/apache/incubator-beam/blob/python-sdk/sdks/python/apache_beam/io/filebasedsource.py#L120

[2] https://github.com/apache/incubator-beam/blob/python-sdk/sdks/java/core/src/main/java/org/apache/beam/sdk/io/FileBasedSource.java#L204




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