[GitHub] [arrow-datafusion] comphead commented on a diff in pull request #6663: Minor: Add more doc strings to WindowExpr

["comphead (via GitHub)" <gi...@apache.org>]

comphead commented on code in PR #6663:
URL: https://github.com/apache/arrow-datafusion/pull/6663#discussion_r1229771349


##########
datafusion/physical-expr/src/window/partition_evaluator.rs:
##########
@@ -130,9 +130,41 @@ pub trait PartitionEvaluator: Debug + Send {
         ))
     }
 
-    /// Called for window functions that *do not use* values from the
-    /// the window frame, such as `ROW_NUMBER`, `RANK`, `DENSE_RANK`,
-    /// `PERCENT_RANK`, `CUME_DIST`, `LEAD`, `LAG`).
+    /// Evaluate a window function on the entire input partition,
+    /// `values`, producing one output row for each input.
+    ///
+    /// This function is called once per input *partition* for window
+    /// functions that *do not use* values from the window frame,
+    /// such as `ROW_NUMBER`, `RANK`, `DENSE_RANK`, `PERCENT_RANK`,
+    /// `CUME_DIST`, `LEAD`, `LAG`).
+    ///
+    /// The function is passed the window as the `value` and must
+    /// produce an output column with one output row for every input
+    /// row.
+    ///
+    /// `num_rows` is requied to correctly compute the output in case
+    /// `values.len() == 0`
+    ///
+    /// Using this function is an optimization: certain window
+    /// functions are not affected by the window frame definition, and
+    /// thus using `evaluate`, DataFusion can skip the (costly) window
+    /// frame boundary calculation.
+    ///
+    /// For example, the `LAG` built in window function does not use
+    /// the values of its window frame (it can be computed in one shot
+    /// on the entire partition with `Self::evalute` regardless of the
+    /// window defined in the `OVER` clause)
+    ///
+    /// ```sql
+    /// lag(x, 1) OVER (ROWS BETWEEN 2 PRECEDING AND 3 FOLLOWING)

Review Comment:
   interesting that lag doesn't require explicit ordering, but it works



-- 
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: github-unsubscribe@arrow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org