Uploaded image for project: 'Cassandra'
  1. Cassandra
  2. CASSANDRA-6292

CQL should be able to do a delete with a > or < operator on a clustering portion of a primary key

    XMLWordPrintableJSON

Details

    • Improvement
    • Status: Resolved
    • Normal
    • Resolution: Duplicate
    • 2.1 beta1
    • Legacy/CQL
    • None

    Description

      The following creates a table that at the storage engine level, has "key" as the row key, and evt as a time ordered list of columns within the row:
      create table test (key text, evt timestamp, primary key(key,evt)) with clustering order by (evt asc);

      Given the ability to create slice tombstones (both implicitly by deleting collections and supercolumns) as well as in thrift through a delete with a slice predicate.
      Given the schema above, the expected syntax to delete all columns with key "one" and timestamp less than 1000 would be:
      delete from test where key='one' and evt < 1000;
      and if you did that as well as
      delete from test where key='one' and evt > 1000;
      You would accomplish exactly what is being asked about (in terms of thrift and astyanax) here:
      http://stackoverflow.com/questions/19747005/delete-all-the-columns-and-its-data-except-for-one-columns-using-astyanax-client

      Attachments

        Issue Links

          Activity

            People

              jbellis Jonathan Ellis
              tupshin Tupshin Harper
              Jonathan Ellis
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: