Correct relation size estimate with low fillfactor

Since commit 29cf61ade3, table_block_relation_estimate_size() considers
fillfactor when estimating number of rows in a relation before the first
ANALYZE. The formula however did not consider tuples may be larger than
available space determined by fillfactor, ending with density 0. This
ultimately means the relation was estimated to contain a single row.

The executor however places at least one tuple per page, even with very
low fillfactor values, so the density should be at least 1. Fixed by
clamping the density estimate using clamp_row_est().

Reported by Heikki Linnakangas. Fix by me, with regression test inspired
by example provided by Heikki.

Backpatch to 17, where the issue was introduced.

Reported-by: Heikki Linnakangas
Backpatch-through: 17
Discussion: https://postgr.es/m/2bf9d973-7789-4937-a7ca-0af9fb49c71e@iki.fi

diff --git a/src/backend/access/table/tableam.c b/src/backend/access/table/tableam.c
index e18a8f8250fd8440b2e8ba49092880f82a0d59dc..a56c5eceb14ade089522679798dcd62509462e3a 100644 (file)
--- a/src/backend/access/table/tableam.c
+++ b/src/backend/access/table/tableam.c
@@ -24,6 +24,7 @@
 #include "access/syncscan.h"
 #include "access/tableam.h"
 #include "access/xact.h"
+#include "optimizer/optimizer.h"
 #include "optimizer/plancat.h"
 #include "port/pg_bitutils.h"
 #include "storage/bufmgr.h"
@@ -740,6 +741,8 @@ table_block_relation_estimate_size(Relation rel, int32 *attr_widths,
        tuple_width += overhead_bytes_per_tuple;
        /* note: integer division is intentional here */
        density = (usable_bytes_per_page * fillfactor / 100) / tuple_width;
+       /* There's at least one row on the page, even with low fillfactor. */
+       density = clamp_row_est(density);
    }
    *tuples = rint(density * (double) curpages);
 

diff --git a/src/test/regress/expected/stats.out b/src/test/regress/expected/stats.out
index 7d91f047bb389cabef8e9db9fb7e9a029685d662..093e6368dbbe1364e226e729b1133a0b0d5d2d1a 100644 (file)
--- a/src/test/regress/expected/stats.out
+++ b/src/test/regress/expected/stats.out
@@ -1749,4 +1749,21 @@ SELECT COUNT(*) FROM brin_hot_3 WHERE a = 2;
 
 DROP TABLE brin_hot_3;
 SET enable_seqscan = on;
+-- Test that estimation of relation size works with tuples wider than the
+-- relation fillfactor. We create a table with wide inline attributes and
+-- low fillfactor, insert rows and then see how many rows EXPLAIN shows
+-- before running analyze. We disable autovacuum so that it does not
+-- interfere with the test.
+CREATE TABLE table_fillfactor (
+  n char(1000)
+) with (fillfactor=10, autovacuum_enabled=off);
+INSERT INTO table_fillfactor
+SELECT 'x' FROM generate_series(1,1000);
+SELECT * FROM check_estimated_rows('SELECT * FROM table_fillfactor');
+ estimated | actual 
+-----------+--------
+      1000 |   1000
+(1 row)
+
+DROP TABLE table_fillfactor;
 -- End of Stats Test

diff --git a/src/test/regress/sql/stats.sql b/src/test/regress/sql/stats.sql
index 11628ebc8a17aeaa734c4701e1f8438ae3ce61e1..0a44e14d9f4a2b2295d06cd82e3162f27507b5f7 100644 (file)
--- a/src/test/regress/sql/stats.sql
+++ b/src/test/regress/sql/stats.sql
@@ -895,4 +895,20 @@ DROP TABLE brin_hot_3;
 
 SET enable_seqscan = on;
 
+-- Test that estimation of relation size works with tuples wider than the
+-- relation fillfactor. We create a table with wide inline attributes and
+-- low fillfactor, insert rows and then see how many rows EXPLAIN shows
+-- before running analyze. We disable autovacuum so that it does not
+-- interfere with the test.
+CREATE TABLE table_fillfactor (
+  n char(1000)
+) with (fillfactor=10, autovacuum_enabled=off);
+
+INSERT INTO table_fillfactor
+SELECT 'x' FROM generate_series(1,1000);
+
+SELECT * FROM check_estimated_rows('SELECT * FROM table_fillfactor');
+
+DROP TABLE table_fillfactor;
+
 -- End of Stats Test