alamb opened a new issue, #10505: URL: https://github.com/apache/datafusion/issues/10505
### Is your feature request related to a problem or challenge? This API is used in several optimizer passes to recursively find all column references in an expression pub fn [Expr::to_columns](https://docs.rs/datafusion/latest/datafusion/logical_expr/enum.Expr.html#method.to_columns)(&self) -> [Result](https://doc.rust-lang.org/nightly/core/result/enum.Result.html)<[HashSet](https://doc.rust-lang.org/nightly/std/collections/hash/set/struct.HashSet.html)<[Column](https://docs.rs/datafusion/latest/datafusion/common/struct.Column.html)>, [DataFusionError](https://docs.rs/datafusion/latest/datafusion/common/enum.DataFusionError.html)> However, it requires `clone()` ing all `Columns` (which requires allocating and copying a string) even if the caller only needs to check if there is a reference. This is inefficient and leads to a slower optimizer ### Describe the solution you'd like I would like an API that does not require copying ### Describe alternatives you've considered One version might be this (returned hasset has `&Expr` not `Expr) ```rust impl Expr { /// Return all referenced `Column`s anywhere in this `Expr` /// /// For example /// * `a+b` would return `{a,b}` /// * `a + 1 + a` would return `{a}` fn column_refs(&self) -> HashSet<&Expr> { ... } ``` ### Additional context _No response_ -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
