thanks for reply, I used Antlr to implement the JOIN operation, if calcite could support join the stream and rpc together, it will easier to achieve the sample build MLOps tasks.
On 2024/01/11 09:06:11 Alessandro Solimando wrote: > Hello, > not sure if it's not possible but the task at hand seems like a good fit > for streaming engines like Apache Flink or Apache Spark. > > Any specific reasons for not using them and trying to achieve this via > Calcite? > > Just for reference, Apache Flink uses Calcite for query planning. > > Best regards, > Alessandro > > On Thu, 11 Jan 2024 at 09:10, 李治 <lx...@gmail.com> wrote: > > > > > > > > 下面是被转发的邮件： > > > > > > 发件人: 李治 <lx...@gmail.com> > > > 主题: [Question] is there any way to join stream table and http request > > table use apache calcite framework? > > > 日期: 2024年1月11日 GMT+8 16:06:49 > > > 收件人: dev-subscr...@calcite.apache.org > > > > > > I would like to join kafka stream table and http request table( > > https://github.com/gaohanghbut/yugo) like this: > > > > > > > > > > > > . my table extend from AbstractQueryableTable.java ,so I implement > > asQueryable interface, but I found http table will be invoke before kafka > > message arrived: > > > > > > > > > > > > inner.toLookup() (HttpQueryable) will convert to lookup object, when > > kafka message arrived, outers.moveNext() will be notified and get data from > > innerLookup which is already put http result in it, like this : > > > > > > > > > > > > maybe calcite can not join stream table with http request? > > > > > > I try to override toLookup interface in my HttpQueryable like this: > > > > > > > > > > > > , but HttpQueryable lookup function dose not invoke by > > EnumerableDefaults.hashEquiJoin_() function, hashEquiJoin_ still call base > > class toLookup interface even I override toLookup in my HttpQueryable: > > > > > > > > > > > > I am so confuse about this, any one can help? > > > > >
