[#20831] YSQL: Show distinct prefix keys explicitly in the explain ou…

…tput

Summary:
### Objective

Before this change, the explain output looks as follows

```
yugabyte=# explain select distinct  col_int_key from mm;
                                            QUERY PLAN
---------------------------------------------------------------------------------------------------
 Unique  (cost=0.00..12.00 rows=93 width=4)
   ->  Distinct Index Only Scan using idx_mm_col_int_key on mm  (cost=0.00..12.00 rows=93 width=4)
         Distinct Prefix: 1
(3 rows)
```

We displayed the length of the distinct prefix because that is the sole parameter necessary for the HybridScan on the DocDB side.

However, users may find it more useful if we instead displayed the keys corresponding to the index columns instead. The new output looks

```
yugabyte=# explain select distinct k1 from t;
                                    QUERY PLAN
----------------------------------------------------------------------------------
 Unique  (cost=0.00..22.80 rows=200 width=4)
   ->  Distinct Index Scan using t_pkey on t  (cost=0.00..22.80 rows=200 width=4)
         Distinct Prefix Keys: k1
(3 rows)
```

### Solution

Pick the prefix length number of leading columns from the list of index targets and print them in explain.
Also, change the regression test output to reflect the change.
Jira: DB-9820

Test Plan:
Jenkins
./yb_build.sh --java-test TestPgRegressDistinctPushdown

Reviewers: tnayak

Reviewed By: tnayak

Subscribers: yql

Differential Revision: https://phorge.dev.yugabyte.com/D33042

src/postgres/src/backend/commands/explain.c

@@ -144,7 +144,8 @@ static void
 YbAggregateExplainableRPCRequestStat(ExplainState			 *es,
 									 const YbInstrumentation *instr);
 static void YbExplainDistinctPrefixLen(
-	int yb_distinct_prefixlen, ExplainState *es);
+	PlanState *planstate, List *indextlist, int yb_distinct_prefixlen,
+	ExplainState *es, List *ancestors);
 static void show_ybtidbitmap_info(YbBitmapTableScanState *planstate,
 								  ExplainState *es);
 
@@ -2413,7 +2414,8 @@ ExplainNode(PlanState *planstate, List *ancestors,
 			 * that's currently the order of operations in DocDB.
 			 */
 			YbExplainDistinctPrefixLen(
-				((IndexScan *) plan)->yb_distinct_prefixlen, es);
+				planstate, ((IndexScan *) plan)->indextlist,
+				((IndexScan *) plan)->yb_distinct_prefixlen, es, ancestors);
 			show_scan_qual(((IndexScan *) plan)->yb_idx_pushdown.quals,
 						   "Storage Index Filter", planstate, ancestors, es);
 			show_scan_qual(((IndexScan *) plan)->yb_rel_pushdown.quals,
@@ -2451,7 +2453,8 @@ ExplainNode(PlanState *planstate, List *ancestors,
 			 * that's currently the order of operations in DocDB.
 			 */
 			YbExplainDistinctPrefixLen(
-				((IndexOnlyScan *) plan)->yb_distinct_prefixlen, es);
+				planstate, ((IndexOnlyScan *) plan)->indextlist,
+				((IndexOnlyScan *) plan)->yb_distinct_prefixlen, es, ancestors);
 			/*
 			 * Storage filter is applied first, so it is output first.
 			 */
@@ -5059,15 +5062,56 @@ YbAggregateExplainableRPCRequestStat(ExplainState			 *es,
  * --------------
  * Distinct Index Scan
  *       ...
- * 	 Distinct Prefix: <prefix length>
+ * 	 Distinct Keys: <Index Prefix Keys>
  *       ...
  *
  * Adds Distinct Prefix to explain info
  */
 static void
-YbExplainDistinctPrefixLen(int yb_distinct_prefixlen, ExplainState *es)
+YbExplainDistinctPrefixLen(PlanState *planstate, List *indextlist,
+						   int yb_distinct_prefixlen, ExplainState *es,
+						   List *ancestors)
 {
 	if (yb_distinct_prefixlen > 0)
-		ExplainPropertyInteger(
-			"Distinct Prefix", NULL, yb_distinct_prefixlen, es);
+	{
+		/* Print distinct prefix keys. */
+		List		   *context;
+		List		   *result = NIL;
+		StringInfoData	distinct_prefix_key_buf;
+		bool			useprefix;
+		int				keyno;
+		ListCell	   *tlelc;
+
+		initStringInfo(&distinct_prefix_key_buf);
+
+		/* Set up deparsing context */
+		context = set_deparse_context_planstate(es->deparse_cxt,
+												(Node *) planstate,
+												ancestors);
+		useprefix = (list_length(es->rtable) > 1 || es->verbose);
+
+		keyno = 0;
+		foreach(tlelc, indextlist)
+		{
+			TargetEntry	*indextle;
+			char 		*exprstr;
+
+			if (keyno >= yb_distinct_prefixlen)
+				break;
+
+			indextle = (TargetEntry *) lfirst(tlelc);
+
+			/* Deparse the expression, showing any top-level cast */
+			exprstr = deparse_expression((Node *) indextle->expr, context,
+										useprefix, true);
+			resetStringInfo(&distinct_prefix_key_buf);
+			appendStringInfoString(&distinct_prefix_key_buf, exprstr);
+			/* Emit one property-list item per key */
+			result = lappend(result, pstrdup(distinct_prefix_key_buf.data));
+
+			keyno++;
+		}
+
+		ExplainPropertyList("Distinct Keys", result, es);
+	}
 }

src/postgres/src/test/regress/expected/yb_aggregates.out

@@ -715,7 +715,7 @@ EXPLAIN (COSTS OFF) SELECT DISTINCT int_8 FROM ybaggtest;
  HashAggregate
    Group Key: int_8
    ->  Distinct Index Only Scan using ybaggtestindex on ybaggtest
-         Distinct Prefix: 2
+         Distinct Keys: int_8, int_2
 (4 rows)
 
 EXPLAIN (COSTS OFF) SELECT COUNT(distinct int_4), SUM(int_4) FROM ybaggtest;

src/postgres/src/test/regress/expected/yb_bitmap_scans_distinct.out

@@ -15,7 +15,7 @@ SELECT DISTINCT r1 FROM test_distinct WHERE r1 < 2 ORDER BY r1;
  Unique (actual rows=1 loops=1)
    ->  Distinct Index Scan using test_distinct_pkey on test_distinct (actual rows=2 loops=1)
          Index Cond: (r1 < 2)
-         Distinct Prefix: 1
+         Distinct Keys: r1
          Storage Table Read Requests: 2
          Storage Table Rows Scanned: 2
 (6 rows)
@@ -32,7 +32,7 @@ SELECT DISTINCT r1, r2 FROM test_distinct WHERE r1 < 2 OR r2 < 3 ORDER BY r1, r2
 ---------------------------------------------------------------------------------------------
  Unique (actual rows=6 loops=1)
    ->  Distinct Index Scan using test_distinct_pkey on test_distinct (actual rows=7 loops=1)
-         Distinct Prefix: 2
+         Distinct Keys: r1, r2
          Storage Filter: ((r1 < 2) OR (r2 < 3))
 (4 rows)
 

src/postgres/src/test/regress/expected/yb_distinct_pushdown_base.out

@@ -31,7 +31,7 @@ EXPLAIN (ANALYZE, COSTS OFF, TIMING OFF, SUMMARY OFF) SELECT DISTINCT r1 FROM t;
 ---------------------------------------------------------------------
  Unique (actual rows=2 loops=1)
    ->  Distinct Index Scan using t_pkey on t (actual rows=3 loops=1)
-         Distinct Prefix: 1
+         Distinct Keys: r1
 (3 rows)
 
 SELECT DISTINCT r1 FROM t;
@@ -47,7 +47,7 @@ EXPLAIN (ANALYZE, COSTS OFF, TIMING OFF, SUMMARY OFF) SELECT DISTINCT r1, r2 FRO
 ---------------------------------------------------------------------
  Unique (actual rows=6 loops=1)
    ->  Distinct Index Scan using t_pkey on t (actual rows=7 loops=1)
-         Distinct Prefix: 2
+         Distinct Keys: r1, r2
 (3 rows)
 
 SELECT DISTINCT r1, r2 FROM t;
@@ -71,7 +71,7 @@ EXPLAIN (ANALYZE, COSTS OFF, TIMING OFF, SUMMARY OFF) SELECT DISTINCT r2 FROM t;
  HashAggregate (actual rows=3 loops=1)
    Group Key: r2
    ->  Distinct Index Scan using t_pkey on t (actual rows=7 loops=1)
-         Distinct Prefix: 2
+         Distinct Keys: r1, r2
 (4 rows)
 
 SELECT DISTINCT r2 FROM t;
@@ -90,7 +90,7 @@ EXPLAIN (ANALYZE, COSTS OFF, TIMING OFF, SUMMARY OFF) SELECT DISTINCT r1 FROM t
  Limit (actual rows=2 loops=1)
    ->  Unique (actual rows=2 loops=1)
          ->  Distinct Index Scan using t_pkey on t (actual rows=3 loops=1)
-               Distinct Prefix: 1
+               Distinct Keys: r1
 (4 rows)
 
 SELECT DISTINCT r1 FROM t LIMIT 2;
@@ -109,7 +109,7 @@ EXPLAIN (ANALYZE, COSTS OFF, TIMING OFF, SUMMARY OFF) SELECT DISTINCT r1 FROM tr
 -----------------------------------------------------------------------
  Unique (actual rows=1 loops=1)
    ->  Distinct Index Scan using tr_pkey on tr (actual rows=1 loops=1)
-         Distinct Prefix: 1
+         Distinct Keys: r1
 (3 rows)
 
 SELECT DISTINCT r1 FROM tr;
@@ -127,7 +127,7 @@ EXPLAIN (ANALYZE, COSTS OFF, TIMING OFF, SUMMARY OFF) SELECT DISTINCT r1 FROM ts
 -----------------------------------------------------------------------
  Unique (actual rows=1 loops=1)
    ->  Distinct Index Scan using ts_pkey on ts (actual rows=1 loops=1)
-         Distinct Prefix: 1
+         Distinct Keys: r1
 (3 rows)
 
 SELECT DISTINCT r1 FROM ts;
@@ -170,7 +170,7 @@ EXPLAIN (ANALYZE, COSTS OFF, TIMING OFF, SUMMARY OFF) SELECT DISTINCT r1 FROM t
 ---------------------------------------------------------------------
  Unique (actual rows=2 loops=1)
    ->  Distinct Index Scan using t_pkey on t (actual rows=3 loops=1)
-         Distinct Prefix: 1
+         Distinct Keys: r1
 (3 rows)
 
 SELECT DISTINCT r1 FROM t ORDER BY r1;
@@ -185,7 +185,7 @@ EXPLAIN (ANALYZE, COSTS OFF, TIMING OFF, SUMMARY OFF) SELECT DISTINCT r1 FROM t
 ------------------------------------------------------------------------------
  Unique (actual rows=2 loops=1)
    ->  Distinct Index Scan Backward using t_pkey on t (actual rows=3 loops=1)
-         Distinct Prefix: 1
+         Distinct Keys: r1
 (3 rows)
 
 SELECT DISTINCT r1 FROM t ORDER BY r1 DESC;
@@ -201,7 +201,7 @@ EXPLAIN (ANALYZE, COSTS OFF, TIMING OFF, SUMMARY OFF) SELECT DISTINCT r1, r2 FRO
 ---------------------------------------------------------------------
  Unique (actual rows=6 loops=1)
    ->  Distinct Index Scan using t_pkey on t (actual rows=7 loops=1)
-         Distinct Prefix: 2
+         Distinct Keys: r1, r2
 (3 rows)
 
 SELECT DISTINCT r1, r2 FROM t ORDER BY r1, r2;
@@ -220,7 +220,7 @@ EXPLAIN (ANALYZE, COSTS OFF, TIMING OFF, SUMMARY OFF) SELECT DISTINCT r1, r2 FRO
 ------------------------------------------------------------------------------
  Unique (actual rows=6 loops=1)
    ->  Distinct Index Scan Backward using t_pkey on t (actual rows=7 loops=1)
-         Distinct Prefix: 2
+         Distinct Keys: r1, r2
 (3 rows)
 
 SELECT DISTINCT r1, r2 FROM t ORDER BY r1 DESC, r2 DESC;
@@ -245,7 +245,7 @@ EXPLAIN (ANALYZE, COSTS OFF, TIMING OFF, SUMMARY OFF) SELECT DISTINCT r1, r2 FRO
    Sort Method: quicksort
    ->  Unique (actual rows=6 loops=1)
          ->  Distinct Index Scan using t_pkey on t (actual rows=7 loops=1)
-               Distinct Prefix: 2
+               Distinct Keys: r1, r2
 (6 rows)
 
 SELECT DISTINCT r1, r2 FROM t ORDER BY r2;
@@ -271,7 +271,7 @@ EXPLAIN (ANALYZE, COSTS OFF, TIMING OFF, SUMMARY OFF) SELECT DISTINCT r1, r2 FRO
 -----------------------------------------------------------------------
  Unique (actual rows=3 loops=1)
    ->  Distinct Index Scan using tm_pkey on tm (actual rows=3 loops=1)
-         Distinct Prefix: 2
+         Distinct Keys: r1, r2
 (3 rows)
 
 SELECT DISTINCT r1, r2 FROM tm;
@@ -288,7 +288,7 @@ EXPLAIN (ANALYZE, COSTS OFF, TIMING OFF, SUMMARY OFF) SELECT DISTINCT r1, r2 FRO
 --------------------------------------------------------------------------------
  Unique (actual rows=3 loops=1)
    ->  Distinct Index Scan Backward using tm_pkey on tm (actual rows=3 loops=1)
-         Distinct Prefix: 2
+         Distinct Keys: r1, r2
 (3 rows)
 
 SELECT DISTINCT r1, r2 FROM tm ORDER BY r1, r2 DESC;