Uploaded image for project: 'Hive'
  1. Hive
  2. HIVE-787

Hive Freeway - support near-realtime data processing

    Details

    • Type: New Feature
    • Status: Resolved
    • Priority: Major
    • Resolution: Won't Fix
    • Affects Version/s: None
    • Fix Version/s: None
    • Component/s: None
    • Labels:
      None

      Description

      Most people are using Hive for daily (or at most hourly) data processing.
      We want to explore what are the obstacles for using Hive for 15 minutes, 5 minutes or even 1 minute data processing intervals, and remove these obstacles.

        Issue Links

          Activity

          Hide
          appodictic Edward Capriolo added a comment -

          What about some scribe like feature where the MAP phase or REDUCE phase can write/append to hive table and partition

          Show
          appodictic Edward Capriolo added a comment - What about some scribe like feature where the MAP phase or REDUCE phase can write/append to hive table and partition
          Hide
          appodictic Edward Capriolo added a comment -

          My previous comment was unclear. Now that Hadoop has append support, it would be nice to be able to write directly from a Hadoop Job doing data ingestion directly to a hive table. In my case, I am pulling files info DFS to be later added with 'add file'. It would be nice if I had a HiveOutputFormat and i could emit() data to hive.

          Show
          appodictic Edward Capriolo added a comment - My previous comment was unclear. Now that Hadoop has append support, it would be nice to be able to write directly from a Hadoop Job doing data ingestion directly to a hive table. In my case, I am pulling files info DFS to be later added with 'add file'. It would be nice if I had a HiveOutputFormat and i could emit() data to hive.
          Hide
          hammer Jeff Hammerbacher added a comment -

          More details on the Data Freeway implementation at Facebook: http://vimeo.com/15337985

          Show
          hammer Jeff Hammerbacher added a comment - More details on the Data Freeway implementation at Facebook: http://vimeo.com/15337985
          Hide
          appodictic Edward Capriolo added a comment -

          The rise of n million stream processing solutions makes it unlikely anyone would attempt to implement this directly. It looks like people are using calcite in real time platforms like Samza so in effect I would say this was done in another way. Reopen if you feel differently.

          Show
          appodictic Edward Capriolo added a comment - The rise of n million stream processing solutions makes it unlikely anyone would attempt to implement this directly. It looks like people are using calcite in real time platforms like Samza so in effect I would say this was done in another way. Reopen if you feel differently.

            People

            • Assignee:
              Unassigned
              Reporter:
              zshao Zheng Shao
            • Votes:
              0 Vote for this issue
              Watchers:
              22 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved:

                Development