Handle no DISTINCT ON expressions and extend the docs for the replace_distinct_aggregate optimizer

Author: gruuya

datafusion/expr/src/logical_plan/plan.rs

@@ -2204,13 +2204,17 @@ pub struct DistinctOn {
 }
 
 impl DistinctOn {
-    /// Create a new `DistintOn` struct.
+    /// Create a new `DistinctOn` struct.
     pub fn try_new(
         on_expr: Vec<Expr>,
         select_expr: Vec<Expr>,
         sort_expr: Option<Vec<Expr>>,
         input: Arc<LogicalPlan>,
     ) -> Result<Self> {
+        if on_expr.is_empty() {
+            return plan_err!("No `ON` expressions provided");
+        }
+
         let on_expr = normalize_cols(on_expr, input.as_ref())?;
 
         // Create fields with any qualifier stuffed in the name itself

datafusion/optimizer/src/replace_distinct_aggregate.rs

@@ -36,6 +36,23 @@ use datafusion_expr::{Aggregate, Distinct, DistinctOn, Expr, LogicalPlan};
 /// ```text
 /// SELECT a, b FROM tab GROUP BY a, b
 /// ```
+///
+/// On the other hand, in the case of a [[DistinctOn]] query the replacement is
+/// a bit more involved and effectively converts
+/// ```text
+/// SELECT DISTINCT ON (a) b FROM tab ORDER BY a DESC, c
+/// ```
+///
+/// into
+/// ```text
+/// SELECT b FROM (
+///     SELECT a, FIRST_VALUE(b ORDER BY a DESC, c) AS b
+///     FROM tab
+///     GROUP BY a
+/// )
+/// ORDER BY a DESC
+/// ```
+/// ```
 
 /// Optimizer that replaces logical [[Distinct]] with a logical [[Aggregate]]
 #[derive(Default)]

datafusion/sqllogictest/test_files/distinct_on.slt

@@ -92,6 +92,10 @@ SELECT DISTINCT ON (c1) c2 FROM aggregate_test_100 ORDER BY c1, c3;
 query error SELECT DISTINCT ON expressions must match initial ORDER BY expressions
 SELECT DISTINCT ON (c2 % 2 = 0) c2, c3 - 100 FROM aggregate_test_100 ORDER BY c2, c3;
 
+# ON expressions are empty
+query error DataFusion error: Error during planning: No `ON` expressions provided
+SELECT DISTINCT ON () c1, c2 FROM aggregate_test_100 ORDER BY c1, c2;
+
 # Use expressions in the ON and ORDER BY clauses, as well as the selection
 query II
 SELECT DISTINCT ON (c2 % 2 = 0) c2, c3 - 100 FROM aggregate_test_100 ORDER BY c2 % 2 = 0, c3 DESC;