[jira] [Created] (ZEPPELIN-1330) hive JDBC SqlCompleter takes very long to column/schema meta info when hive has a lot of tables

["Qiuzhuang Lian (JIRA)" <ji...@apache.org>]

Qiuzhuang Lian created ZEPPELIN-1330:
----------------------------------------

             Summary: hive JDBC SqlCompleter takes very long to column/schema meta info when hive has a lot of tables
                 Key: ZEPPELIN-1330
                 URL: https://issues.apache.org/jira/browse/ZEPPELIN-1330
             Project: Zeppelin
          Issue Type: Bug
          Components: interpreter-auto-completion, Interpreters
    Affects Versions: 0.7.0
            Reporter: Qiuzhuang Lian


When use %hive, it takes very long(at least 3 mins) to connect to HS2.  As check HS2 log, we see it loads a lot of hive schema/column meta which takes long time to complete as in SqlCompleter code as follows:

  public static Set<String> getDataModelMetadataCompletions(Connection connection)
      throws SQLException {
    Set<String> completions = new TreeSet<>();
    if (null != connection) {
      getColumnNames(connection.getMetaData(), completions);
      getSchemaNames(connection.getMetaData(), completions);
    }
    return completions;
  }

This class seems to be for web auto-completion. However, I don't see this feature in web instead BTW.





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