Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-22330

Impact of field names

    XMLWordPrintableJSON

Details

    • Icon: Question Question
    • Resolution: Done
    • Icon: Major - P3 Major - P3
    • None
    • None
    • WiredTiger
    • None
    • Linux 2.6.32-504.3.3.el6.x86_64 #1 SMP x86_64 x86_64 x86_64 GNU/Linux

    Description

      We are mapping one type of our existing in-memory objects to mongodb. All field names are predefined and they have been used in production.

      I did some simple experiment - if the field is an indexed field, the long field name seems to have a big performance (throughput) impact.

      If the field is not an indexed field, the long field name seems to have very little impact, performance wise or storage wise.

      I am using wiredTiger which by default has compression enabled.
      My test is a batch loading app.

      Some mongodb blog mentions that wiredTiger compression will make the long field names less of an issue for storage and for performance.
      Since based on my experiment, that's not the case. I would like to have a better understanding on how long field names impact performance:
      1. Is there any configuration recommendations for handling repeated long field names?
      2. What is the performance and storage (memory) impact of long indexed field names?
      3. What is the performance and storage impact of long non-indexed field names?

      Thanks in advance!
      Judy

      Attachments

        Activity

          People

            Unassigned Unassigned
            Judy.Han Judy Han [X]
            Votes:
            0 Vote for this issue
            Watchers:
            6 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: