Hi Henry, If you have converted the mysql table to a flink stream table. In flink table/sql, streams and stream joins can also do this, such as setting the state retention time of one of the tables to be permanent. But when the job is just running, you may not be able to match the results, because the data belonging to the mysql table is just beginning to play as a stream.
Thanks, vino. 徐涛 <happydexu...@gmail.com> 于2018年9月25日周二 下午5:10写道： > Hi Vino & Hequn, > I am now using the table/sql API, if I import the mysql table as a stream > then convert it into a table, it seems that it can also be a workaround for > batch/streaming joining. May I ask what is the difference between the UDTF > method? Does this implementation has some defects? > Best > Henry > > 在 2018年9月22日，上午10:28，Hequn Cheng <chenghe...@gmail.com> 写道： > > Hi > > +1 for vino's answer. > Also, this kind of join will be supported in FLINK-9712 > <https://issues.apache.org/jira/browse/FLINK-9712>. You can check more > details in the jira. > > Best, Hequn > > On Fri, Sep 21, 2018 at 4:51 PM vino yang <yanghua1...@gmail.com> wrote: > >> Hi Henry, >> >> There are three ways I can think of: >> >> 1) use DataStream API, implement a flatmap UDF to access dimension table; >> 2) use table/sql API, implement a UDTF to access dimension table; >> 3) customize the table/sql join API/statement's implementation (and >> change the physical plan) >> >> Thanks, vino. >> >> 徐涛 <happydexu...@gmail.com> 于2018年9月21日周五 下午4:43写道： >> >>> Hi All, >>> Sometimes some “dimension table” need to be joined from the >>> "fact table", if data are not joined before sent to Kafka. >>> So if the data are joined in Flink, does the “dimension table” >>> have to be import as a stream, or there are some other ways can achieve it? >>> Thanks a lot! >>> >>> Best >>> Henry >> >> >
