Release 1.1.0 (30 Dec 2020)

Key Features

AreaFeaturePR #s
Heuristic IndexBtree index – BTree index is used for split filtering, and is used only by the coordinator nodes. If an index is created on a column which is part of a predicate in the query, then openLooKeng may be able to improve the performance of the query by filtering out splits during scheduling time.392,437,452,457
The bitmap index is used for filtering data read from ORC files, and is only used by the worker nodes. The bitmap index should be created on high-cardinality columns. In this release, the bitmap index is enhanced by storing the keys in a BTree for faster key lookup, index creation, and smaller index sizes.437,418,435,407,390
High performanceSeveral performance improvements where made such as:
· Improved the performance of the window function.
· Changed the dynamic filter wait time to be based on the time the task was scheduled on the worker, and to not include the time spent in planning.
· Apply dynamic filters to semi-joins.
· Implemented a rule to convert left join to semi-join.
· Implemented a rule to convert the self-join to a group by.
Having many small splits affects the scheduling in openLooKeng as it need to do time slice to multiple splits. This causes more scheduling overhead as more times goes in switching across splits and less work done and the wait time of each split to be read also increases due to this. In order to avoid this, Merge Small splits feature is used to group the small splits and schedule together as though it’s a single split scheduling387
Reuse exchange feature introduces a query optimizer to reduce the query latency by caching table data in memoery when it is used more than once in a query with the same projections and filters applied to it443
ConnectorSupport openGauss data source492
Support Mongodb data source464
Refactor ES connector to be compatible with ES 7.X380
Introduce the document for Oracle connector, and enhance Oracle connector to support synonyms.380,487
Enhance the hive connector to support impersonation in Hive Metastore.454
Improve the SQL syntax compatibility of ODBC driver.15,10,19,12,21,22
DC connectorSupport dynamic filter push down in cross data center connector402,419
User experienceProvide a brand new admin web UI with SQL editor and system monitor163,368,404
Simplify the configurations397,369,449
SecuritySupport integration with Apache Ranger for unified permission control491

Known Issues

CategoryDescriptionGitee issue
Data Center ConnectorIf the Data center has many more than 3 tiers, and when the Join condition is on a small value, enabling dynamic filtering takes longer time than disabling dynamic filtering. This is because the column data is not high-cardinality, we need a better filter.I2BAZZ
Heuristic IndexAfter drop index ,the index files folders would not be removed from HDFS. However the files inside them are removedI2BB1N
Creating Index with or-predicate shows the incorrect unsupported error message.I2BB3O
The splits cannot be filtered after create BTree index for transactional partition tableI2BB6M
WEB UIThe response time of the metadata tree list is long after the refresh button is clicked.This is because the cache is invalidated and the data is re-fetched from all catalogs and schemas for all tables and columns.I2BB2B
The text is overflow out from the text area when there are too much live plans information.I2BB4E
InstallationBy using the auto deploy script, the system reused the old configuration file without any prompt messages, this will mislead user.I2BB52
VacuumWhen a vacuum is run more than once on a table which had delete/update operations performed on it, the hive statistics might get corrupted and the queries would fail. To avoid this, set the session flag hive.collect_column_statistics_on_write to false (set session hive.collect_column_statistics_on_write=false) before running VACUUM command. In case the command was run without setting the above flag, run ANALYZE table name command to rectify the statistics. This issue is rectified in PR 517, but not merged into 1.1.0 release.I2BFH9
Reuse ExchangeWhen reuse exchange feature is enabled in (reuse_table_scan=true), querying non hive catalogs would fail. It is recommended use (set session reuse_table_scan=true;) when querying hive catalogs and disable it for other catalogs. This issue is rectified in PR 516, but not merged into 1.1.0 release.I2BEWV

Obtaining the Document

For details, see






● 拼写,格式,无效链接等错误;

● 技术原理、功能、规格等描述和软件不一致,存在错误;

● 原理图、架构图等存在错误;

● 版本号不匹配:文档版本或内容描述和实际软件不一致;

● 对重要数据或系统存在风险的操作,缺少安全提示;

● 排版不美观,影响阅读;


● 描述存在歧义;

● 图形、表格、文字等晦涩难懂;

● 逻辑不清晰,该分类、分项、分步骤的没有给出;


● 很难通过搜索引擎,openLooKeng官网,相关博客找到所需内容;


● 命令、命令参数等错误;

● 命令无法执行或无法完成对应功能;


● 关键步骤错误或缺失,无法指导用户完成任务,比如安装、配置、部署等;

● 逻辑不清晰,该分类、分项、分步骤的没有给出

● 图形、表格、文字等晦涩难懂

● 缺少必要的前提条件、注意事项等;

● 描述存在歧义