Cherry pick Rows out in EXPLAIN ANALYZE

src/backend/cdb/cdbvars.c

@@ -283,6 +283,7 @@ int			gp_hashagg_groups_per_bucket = 5;
 int			gp_motion_slice_noop = 0;
 
 /* Cloudberry Database Experimental Feature GUCs */
+bool		gp_enable_explain_rows_out = false;
 bool		gp_enable_explain_allstat = false;
 bool		gp_enable_motion_deadlock_sanity = false;	/* planning time sanity
 														 * check */

src/backend/commands/explain_gp.c

@@ -941,7 +941,7 @@ cdbexplain_collectStatsFromNode(PlanState *planstate, CdbExplain_SendStatCtx *ct
  */
 typedef struct CdbExplain_DepStatAcc
 {
-	/* vmax, vsum, vcnt, segmax */
+	/* vmax, vmin, vsum, vcnt, segmax, segmin */
 	CdbExplain_Agg agg;
 	/* max's received StatHdr */
 	CdbExplain_StatHdr *rshmax;
@@ -1716,6 +1716,46 @@ cdbexplain_showExecStats(struct PlanState *planstate, ExplainState *es)
 	}
 	pfree(extraData.data);
 
+	/*
+	 * Print "Rows out"
+	 */
+
+	if (gp_enable_explain_rows_out && es->analyze && ns->ninst > 0) {
+        double ntuples_max = ns->ntuples.vmax;
+        int ntuples_imax = ns->ntuples.imax;
+        double ntuples_min = ns->ntuples.vmin;
+        int ntuples_imin = ns->ntuples.imin;
+        double ntuples_avg = cdbexplain_agg_avg(&ns->ntuples);
+        int ntuples_cnt = ns->ntuples.vcnt;
+
+        if (es->format == EXPLAIN_FORMAT_TEXT)
+        {
+            /*
+             * create a header for all stats: separate each individual stat by an
+             * underscore, separate the grouped stats for each node by a slash
+             */
+            appendStringInfoSpaces(es->str, es->indent * 2);
+            appendStringInfoString(es->str, "Rows out: ");
+
+            appendStringInfo(es->str,
+                                 "%.2f rows avg x %d workers, %.0f rows max (seg%d), %.0f rows min (seg%d).\n",
+                                 ntuples_avg,
+                                 ntuples_cnt,
+                                 ntuples_max,
+                                 ntuples_imax,
+                                 ntuples_min,
+                                 ntuples_imin);
+        }
+        else {
+            ExplainPropertyInteger("Workers", NULL, ntuples_cnt, es);
+            ExplainPropertyFloat("Average Rows", NULL, ntuples_avg, 1, es);
+            ExplainPropertyFloat("Max Rows", NULL, ntuples_max, 0, es);
+            ExplainPropertyInteger("Max Rows Segment", NULL, ntuples_imax, es);
+            ExplainPropertyFloat("Min Rows", NULL, ntuples_min, 0, es);
+            ExplainPropertyInteger("Min Rows Segment", NULL, ntuples_imin, es);
+        }
+    }
+
 	/*
 	 * Dump stats for all workers.
 	 */

src/backend/utils/misc/guc_gp.c

@@ -775,6 +775,17 @@ struct config_bool ConfigureNamesBool_gp[] =
 		NULL, NULL, NULL
 	},
 
+	{
+		{"gp_enable_explain_rows_out", PGC_USERSET, CLIENT_CONN_OTHER,
+			gettext_noop("Print avg, min and max rows out and which segments reach them in EXPLAIN ANALYZE."),
+			NULL,
+			GUC_NO_SHOW_ALL | GUC_NOT_IN_SAMPLE
+		},
+		&gp_enable_explain_rows_out,
+		false,
+		NULL, NULL, NULL
+	},
+
 	{
 		{"gp_enable_explain_allstat", PGC_USERSET, CLIENT_CONN_OTHER,
 			gettext_noop("Experimental feature: dump stats for all segments in EXPLAIN ANALYZE."),

src/include/cdb/cdbexplain.h

@@ -26,18 +26,22 @@ struct CdbExplain_ShowStatCtx;          /* private, in "cdb/cdbexplain.c" */
 typedef struct
 {
 	double		vmax;			/* maximum value of statistic */
+    double		vmin;			/* minimum value of statistic */
 	double		vsum;			/* sum of values */
 	int			vcnt;			/* count of values > 0 */
 	int			imax;			/* id of 1st observation having maximum value */
+    int			imin;			/* id of 1st observation having minimum value */
 } CdbExplain_Agg;
 
 static inline void
 cdbexplain_agg_init0(CdbExplain_Agg *agg)
 {
     agg->vmax = 0;
+    agg->vmin = 0;
     agg->vsum = 0;
     agg->vcnt = 0;
     agg->imax = 0;
+    agg->imin = 0;
 }
 
 static inline bool
@@ -48,6 +52,13 @@ cdbexplain_agg_upd(CdbExplain_Agg *agg, double v, int id)
         agg->vsum += v;
         agg->vcnt++;
 
+        if (v < agg->vmin ||
+            agg->vcnt == 1)
+        {
+            agg->vmin = v;
+            agg->imin = id;
+        }
+
         if (v > agg->vmax ||
             agg->vcnt == 1)
         {

src/include/cdb/cdbvars.h

@@ -630,6 +630,12 @@ extern bool gp_enable_agg_pushdown;
  */
 extern bool gp_enable_preunique;
 
+/* May Cloudberry print statistics as average, minimum and maximum rows out
+ * and on which segments reach them for each node during EXPLAIN ANALYZE?
+ *
+ */
+extern bool gp_enable_explain_rows_out;
+
 /* May Cloudberry dump statistics for all segments as a huge ugly string
  * during EXPLAIN ANALYZE?
  *

src/include/utils/unsync_guc_name.h

@@ -182,6 +182,7 @@
 		"gp_enable_agg_pushdown",
 		"gp_enable_ao_indexscan",
 		"gp_enable_direct_dispatch",
+		"gp_enable_explain_rows_out",
 		"gp_enable_explain_allstat",
 		"gp_enable_fast_sri",
 		"gp_enable_global_deadlock_detector",

src/test/regress/expected/gp_explain.out

@@ -436,6 +436,22 @@ explain analyze SELECT * FROM explaintest;
 (8 rows)
 
 set gp_enable_explain_allstat=DEFAULT;
+-- Test explain rows out.
+set gp_enable_explain_rows_out=on;
+explain analyze SELECT * FROM explaintest;
+                                                          QUERY PLAN                                                           
+-------------------------------------------------------------------------------------------------------------------------------
+ Gather Motion 3:1  (slice1; segments: 3)  (cost=0.00..3.10 rows=10 width=4) (actual time=0.266..0.384 rows=5 loops=2)
+   ->  Seq Scan on explaintest  (cost=0.00..3.10 rows=4 width=4) (actual time=0.011..0.013 rows=2 loops=2)
+         Rows out: 3.33 rows avg x 3 workers, 5 rows max (seg1), 2 rows min (seg2).
+   (slice0)    Executor memory: 322K bytes.
+   (slice1)    Executor memory: 50K bytes avg x 3 workers, 50K bytes max (seg0).
+ Memory used:  128000kB
+ Optimizer: Postgres query optimizer
+ Total runtime: 2.600 ms
+(8 rows)
+
+set gp_enable_explain_rows_out=DEFAULT;
 --
 -- Test GPDB-specific EXPLAIN (SLICETABLE) option.
 --

src/test/regress/expected/gp_explain_optimizer.out

@@ -458,6 +458,22 @@ explain analyze SELECT * FROM explaintest;
 (8 rows)
 
 set gp_enable_explain_allstat=DEFAULT;
+-- Test explain rows out.
+set gp_enable_explain_rows_out=on;
+explain analyze SELECT * FROM explaintest;
+                                                          QUERY PLAN                                                           
+-------------------------------------------------------------------------------------------------------------------------------
+ Gather Motion 3:1  (slice1; segments: 3)  (cost=0.00..431.00 rows=10 width=4) (actual time=0.298..0.302 rows=5 loops=2)
+   ->  Seq Scan on explaintest  (cost=0.00..431.00 rows=4 width=4) (actual time=0.013..0.015 rows=2 loops=2)
+         Rows out: 3.33 rows avg x 3 workers, 5 rows max (seg1), 2 rows min (seg2).
+   (slice0)    Executor memory: 290K bytes.
+   (slice1)    Executor memory: 50K bytes avg x 3 workers, 50K bytes max (seg0).
+ Memory used:  128000kB
+ Optimizer: Pivotal Optimizer (GPORCA) version 3.2.0
+ Total runtime: 1.577 ms
+(8 rows)
+
+set gp_enable_explain_rows_out=DEFAULT;
 --
 -- Test GPDB-specific EXPLAIN (SLICETABLE) option.
 --

src/test/regress/sql/gp_explain.sql

@@ -228,6 +228,11 @@ set gp_enable_explain_allstat=on;
 explain analyze SELECT * FROM explaintest;
 set gp_enable_explain_allstat=DEFAULT;
 
+-- Test explain rows out.
+set gp_enable_explain_rows_out=on;
+explain analyze SELECT * FROM explaintest;
+set gp_enable_explain_rows_out=DEFAULT;
+
 
 --
 -- Test GPDB-specific EXPLAIN (SLICETABLE) option.