Re: estimation problems for DISTINCT ON with FDW

[Tom Lane]

Concretely, I now propose the attached, which seems entirely
safe to back-patch.

                        regards, tom lane


diff --git a/doc/src/sgml/fdwhandler.sgml b/doc/src/sgml/fdwhandler.sgml
index 6587678af2..1e997c218b 100644
--- a/doc/src/sgml/fdwhandler.sgml
+++ b/doc/src/sgml/fdwhandler.sgml
@@ -123,6 +123,14 @@ GetForeignRelSize(PlannerInfo *root,
      should be replaced if at all possible.  The function may also choose to
      update <literal>baserel-&gt;width</literal> if it can compute a better estimate
      of the average result row width.
+     (The initial value is based on column data types and on column
+     average-width values measured by the last <command>ANALYZE</command>.)
+     Also, this function may update <literal>baserel-&gt;tuples</literal> if
+     it can compute a better estimate of the foreign table's total row count.
+     (The initial value is
+     from <structname>pg_class</structname>.<structfield>reltuples</structfield>
+     which represents the total row count seen by the
+     last <command>ANALYZE</command>.)
     </para>
 
     <para>
diff --git a/src/backend/optimizer/path/allpaths.c b/src/backend/optimizer/path/allpaths.c
index d984da25d7..63761d5593 100644
--- a/src/backend/optimizer/path/allpaths.c
+++ b/src/backend/optimizer/path/allpaths.c
@@ -911,6 +911,9 @@ set_foreign_size(PlannerInfo *root, RelOptInfo *rel, RangeTblEntry *rte)
 
 	/* ... but do not let it set the rows estimate to zero */
 	rel->rows = clamp_row_est(rel->rows);
+
+	/* also, make sure rel->tuples is not insane relative to rel->rows */
+	rel->tuples = Max(rel->tuples, rel->rows);
 }
 
 /*