HI, > about the question can I assume that ResolvedCatalogTable will be always a > runtime type.
Sorry for I don't really understand your question , why do you have such assumtion? Best regards, Yuxia 发件人: "Krzysztof Chmielewski" <krzysiek.chmielew...@gmail.com> 收件人: "User" <user@flink.apache.org> 发送时间: 星期六, 2023年 1 月 21日 上午 3:13:12 主题: Re: Custom catalog implementation - getting table schema for computed columns Ok, so now I see that runtime type of "table" parameter is ResolvedCatalogTable that has method getResolvedSchema. So I guess my question is, can I assume that ResolvedCatalogTable will be always a runtime type? pt., 20 sty 2023 o 19:27 Krzysztof Chmielewski < [ mailto:krzysiek.chmielew...@gmail.com | krzysiek.chmielew...@gmail.com ] > napisał(a): Hi, I'm implementing a custom Catalog where for "create table" I need to get tables schema, both column names and types from DDL. Now the Catalog's createTable method has "CatalogBaseTable table" argument. The CatalogBaseTable has a deprecated "getSchema" and suggest to use getUnresolvedSchema instead. I was able to resolve schema types for physical columns, but I'm struggling with computed columns [1]. To be more precise I'm struggling to get//resolve the type of this field. I see that all implementations that would be needed to resolve inderlying expression of UnresolvedComputedColumn are marked as @Internal. On the other hand the deprecated "getSchema" has a proper type for this ComputedColumn. I'm wondering now, what should I do. Should I use the deprecated API that has what I need already or should I use suggested API and somehow try to resolve the type using @Internal classes which also does not seems safe. I would appreciate for any hint here. [1] [ https://nightlies.apache.org/flink/flink-docs-release-1.16/docs/dev/table/sql/create/#:~:text=BIGINT%2C%20%60name%60%20STRING)-,Computed%20Columns,-Computed%20columns%20are | https://nightlies.apache.org/flink/flink-docs-release-1.16/docs/dev/table/sql/create/#:~:text=BIGINT%2C%20%60name%60%20STRING)-,Computed%20Columns,-Computed%20columns%20are ]
