Release Note 2.0.5

[xiaokang]

Previous Release Note 2.0.4

Thanks to our community users and developers, about 217 improvements and bug fixes have been made in Doris 2.0.5 version.

1 Behavior change

• change char function behaviour: select char(0) = '\0' return true as MySQL
  • [function](char) change char function behaviour same with mysql #30034
• Allow exporting empty data
  • [fix](outfile) Fix unable to export empty data #30703

2 New features

• Eliminate left outer join with is null condition
• Add show-tablets-belong stmt for analyzing a batch of tablet-ids
• InferPredicates support In, such as a = b & a in [1, 2] -> b in [1, 2]
• Optimize plan when column stats are unavailable
• Optimize plan using rollup column stats
• Support analyze materialized view
• Support ShowProcessStmt Show all Fe connection

3 Improvement and optimizations

• Optimize query plan when column stats are unaviable
• Optimize query plan using rollup column stats
• Stop analyze quickly after user close auto analyze
• Catch load column stats exception, avoid print too much stack info to fe.out
• Select materialized view by specify the view name in sql
• Change auto analyze max table width default value to 100
• Escape characters for columns in recovery predicate pushdown in jdbc catalog
• Fix jdbc mysql catalog to_date fun pushdown
• Optimize the close logic of JDBC client
• Optimize jdbc connection pool parameter settings
• Obtain hudi partition information through HMS's API
• optimize routine load job error msg and memory
• Skip all backup/restore jobs if max allowd option is set to 0

See the complete list of improvements and bug fixes on github .

Big Thanks

Thanks all who contribute to this release:
@airborne12
@alexxing662
@amorynan
@AshinGau
@BePPPower
@bingquanzhao
@BiteTheDDDDt
@ByteYue
@caiconghui
@cambyzju
@catpineapple
@dataroaring
@eldenmoon
@Emor-nj
@englefly
@felixwluo
@GoGoWen
@HappenLee
@hello-stephen
@HHoflittlefish777
@HowardQin
@JackDrogon
@jacktengg
@jackwener
@Jibing-Li
@KassieZ
@LemonLiTree
@liaoxin01
@liugddx
@LuGuangming
@morningman
@morrySnow
@mrhhsg
@Mryange
@mymeiyi
@nextdreamblue
@qidaye
@ryanzryu
@seawinde
@starocean999
@TangSiyang2001
@vinlee19
@w41ter
@wangbo
@wsjz
@wuwenchi
@xiaokang
@XieJiann
@xingyingone
@xy720
@xzj7019
@yujun777
@zclllyybb
@zhangstar333
@zhannngchen
@zhiqiang-hhhh
@zxealous
@zy-kkk
@zzzxl1993

[ryanzryu]

您的邮件我已收到。

[xiaokang]

中文版release note如下：

1 行为变更

• select char(0) = '\0' 返回true，跟MySQL的行为保持一致
  • [function](char) change char function behaviour same with mysql #30034
• export导出数据支持空表
  • [fix](outfile) Fix unable to export empty data #30703

2 新功能

• 利用过滤条件中的is null谓词，将outer join转换为anti join
• 增加SHOW TABLETS BELONG语法用于获取tablet属于哪个table
• InferPredicates 支持IN，例如：a = b & a in [1, 2] -> b in [1, 2]
• 支持对物化视图收集统计信息
• SHOW PROCESSLIST支持输出连接对应的fe
• export导出csv文件支持通过 with_bom 参数控制是否带有Windows BOM

3 改进和优化

• 在无统计信息时优化query plan
• 基于rollup的统计信息优化query plan
• 用户停止auto analyze后尽快停止统计信息收集任务
• 缓存统计信息收集异常，避免大约太多异常栈
• 支持在SQL中自定使用某个物化视图
• JDBC catalog 谓词下推列名字符转义
• 修复MySQL catalog中to_date函数下推的问题
• 优化JDBC客户端连接关闭的逻辑，在异常时正常取消查询
• 优化JDBC连接池的参数
• 通过HMS API获取hudi外表的分区信息
• 优化routine load的内存占用和错误信息
• 如果max_backup_restore_job_num_per_db参数为0，跳过所有备份恢复任务