Re: [DISCUSS] FLIP-154: SQL Implicit Type Coercion

2021-03-14 Thread Jingsong Li
Thanks Danny for your update, the FLIP page looks very good now. Flink already has some implicit casting rules in LogicalTypeCasts[1]. The biggest difference is that the previous implicit casting is strictly limited to the safety, "Implicit casts are used for safe type widening and type generaliza

Re: [DISCUSS] FLIP-154: SQL Implicit Type Coercion

2021-03-10 Thread Danny Chan
Thanks for the feedback, Jingsong ~ This design mainly follows the behaviors of PostgreSQL and SQL-SERVER, because their rules are more in line with the SQL standard. I have fixed the WIKI and add more details about the diff in it. Best, Danny Chan Jingsong Li 于2021年3月11日周四 下午1:34写道: > Thanks

Re: [DISCUSS] FLIP-154: SQL Implicit Type Coercion

2021-03-10 Thread Jingsong Li
Thanks Danny for starting this discussion. Big +1 for Implicit Type Coercion, in my opinion, it is very useful for writing SQL conveniently. I think there are two orthogonal things to discuss here: 1.[Matrix] Which types and which types can be implicitly converted. 2.[Strategies] In different cas

[DISCUSS] FLIP-154: SQL Implicit Type Coercion

2021-03-09 Thread Danny Chan
Hello, fellows, long time no see ~ Here i want to fire a discussion about the SQL implicit type coercion, a required feature for SQL production but missed for Flink SQL for a long time ~ SQL implicit type coercion is very useful for these cases: - the inter-operation within builtin SQL operators