Hyukjin Kwon created SPARK-61600:
------------------------------------

             Summary: Take columns in agg, sort, order_by and the other 
column-list methods of the Rust API, as select and group_by do
                 Key: SPARK-61600
                 URL: https://issues.apache.org/jira/browse/SPARK-61600
             Project: Spark
          Issue Type: Improvement
          Components: Connect
    Affects Versions: connect-rust-4.3.0
            Reporter: Hyukjin Kwon


In the Rust API, DataFrame::select and group_by take any columns (impl 
IntoIterator<Item = impl Into<Column>>), while sort, order_by, 
sort_within_partitions, agg (also of GroupedData), observe, observe_with, 
hint_expressions, repartition_by_range(_opt), repartition_by_expressions and 
WindowSpec/Window::partition_by took a Vec<Expression>, and 
WindowSpec/Window::order_by a Vec<SortOrder>. So the docs and examples wrote:

    df.group_by([f::col("category")]).agg(vec![
        f::sum(f::col("id")).alias("total").expression().clone(),
        f::avg(f::col("id")).alias("average").expression().clone(),
    ])

They should take columns as select and group_by do, with From<Expression> and 
From<SortOrder> for Column, so that callers that pass expressions or sort 
orders still compile.




--
This message was sent by Atlassian Jira
(v10.3.26#10030026)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to