GitHub user dhalperi opened a pull request: https://github.com/apache/beam/pull/2998
[BEAM-1494] Correctly handle content-encoding in GcsFileSystem, fixing reading of such files in CompressedSource R: @jkff thoughts? CC: @chamikaramj You can merge this pull request into a Git repository by running: $ git pull https://github.com/dhalperi/beam b1494-gcs-content-encoding Alternatively you can review and apply these changes as the patch at: https://github.com/apache/beam/pull/2998.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #2998 ---- commit 7ef0f8afc88b292724228fb3507e6d0c77c0b1aa Author: Dan Halperin <dhalp...@google.com> Date: 2017-05-09T19:34:04Z FileBasedSource: isSplittable should not throw This is a legacy design from Dataflow 1.x that was a poor choice. All the information needed to know whether a source is splittable should be known at source construction time, and if runtime behavior is needed it should result in conservative choices, aka false. commit 59e8e0ec27dfc498dacaaf425548681ed07a2d31 Author: Dan Halperin <dhalp...@google.com> Date: 2017-05-09T19:36:10Z CompressedSource: only use delegate reader if the file is splittable Otherwise, it's likely compressed commit b71f5dfed5b8e56dd01cca5a71e2fa72233ab363 Author: Dan Halperin <dhalp...@google.com> Date: 2017-05-09T19:36:53Z GcsFileSystem: mark content-encoded files as not seekable That is the truth (since they are actually compressed) and will result in correct data when reading from them in, e.g., TextIO ---- --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---