Hi friend, I think it is a good question. When related to RDBMS soure, make a correct decision is not so easy, I think it should be depend you requirement and your understanding about these feature. Firstly, source type = 8, is older, it use sqoop to ingest data from RDBMS into hive, thus make "create flat table" realizable. It has been tested under MySQL and SqlServer. After cube built successfully, your cube could answer all measure type provided by Kylin, and it is quicker than query directly aganist underlying RDBMS soure, . Secondly, source type = 16(we called DataSource SDK), is newer. Not only sqoop, but also apache calcite, has been introduced into Kylin, to make Kylin even stronger. In addition to the ability of ingesting data from RDBMS, but also DataSource SDK can rewrite your query to make more query could be answered bu Kylin. In some scenario, when your query cannot be match with your cube, kylin will try pushdown to let real source answer such query. But pushdown may failed bacause reason of difference of RDBMS's dialect. DataSource SDK can make a difference by rewriting your query based on a mapping file(a XML file) provided by you. And it has provided some SPI(service provided interface) to let your implement your rewrite logic, which is more stronger and also more difficult. You should check source code carefully to know how it should be used. So, as far as I can see, source type = 8 is OK for most scenario, it is out of box and should works well. On the another hand,if you want to make push down query more smoothly, you should try source type = 16. If you have more question, or I have make any mistake, please let me know. Thank you very much.
----------------- ----------------- Best wishes to you ! From ：Xiaoxiang Yu At 2019-07-04 20:11:10, "紫电_恶魔" <[email protected]> wrote: The following content is translated from Chinese using translation software, it is inevitable that errors, please treat Kylin.properties Kylin.source.default in the official documentation using jdbc source has 8 and 16 configurations, I do not know the difference http://kylin.apache.org/cn/development/datasource_sdk.html Based on this document, create a new configuration file, postgresql.xml, as follows The SQL executed when building cube is as follows Please tell me what other configuration needs to be done. Thank you
