imay commented on issue #4051: URL: https://github.com/apache/incubator-doris/issues/4051#issuecomment-666881045
> 1. Currently Doris is implemented based on the `Merge on Read` method. This implementation method is the same as the current implementation strategy. Why not change it to 'Merge on Write'? > 2. In principle, users do not need to provide a value column, this function can support users not to specify a value column. One purpose of this function is to receive the change data sent from the TP database, which is usually the entire line of content. I see that this function support all the load interface. And insert data and delete data in one batch, how can user delete data without value column? > 3. This function is only for the table of the unique key model, and has nothing to do with the materialized view How about "AGG_KEYS 表 目前AGG_KEYS 表读取时是从低版本向高版本合并数据，当发现有当前行有delete标记时清除之前的数据，剩下的数据继续合并，如果没有数据跳过读取下一个key。没有发现delete 标记则按现有流程调用agg_update。" Even this function is only for unique model, however this is a "user-interface" level function. We should discuss from a more macroscopic and creative perspective, not the immediate function. If this function is only for unique model, what's the funtion about duplicate and aggregate model? ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org For additional commands, e-mail: commits-h...@doris.apache.org
