Introduce AQO v.1.4. Add reliability field into the aqo_data table.

Author: a.pervushina

Makefile

@@ -1,7 +1,7 @@
 # contrib/aqo/Makefile
 
 EXTENSION = aqo
-EXTVERSION = 1.2
+EXTVERSION = 1.4
 PGFILEDESC = "AQO - Adaptive Query Optimization"
 MODULE_big = aqo
 OBJS = aqo.o auto_tuning.o cardinality_estimation.o cardinality_hooks.o \
@@ -32,7 +32,7 @@ EXTRA_REGRESS_OPTS=--temp-config=$(top_srcdir)/$(subdir)/conf.add
 EXTRA_INSTALL = contrib/postgres_fdw contrib/pg_stat_statements
 
 DATA = aqo--1.0.sql aqo--1.0--1.1.sql aqo--1.1--1.2.sql aqo--1.2.sql \
-		aqo--1.2--1.3.sql
+		aqo--1.2--1.3.sql aqo--1.3--1.4.sql
 
 ifdef USE_PGXS
 PG_CONFIG ?= pg_config

aqo--1.3--1.4.sql

@@ -0,0 +1,6 @@
+/* contrib/aqo/aqo--1.3--1.4.sql */
+
+-- complain if script is sourced in psql, rather than via CREATE EXTENSION
+\echo Use "ALTER EXTENSION aqo UPDATE TO '1.4'" to load this file. \quit
+
+ALTER TABLE public.aqo_data ADD COLUMN reliability double precision [];

aqo.control

@@ -1,5 +1,5 @@
 # AQO extension
 comment = 'machine learning for cardinality estimation in optimizer'
-default_version = '1.3'
+default_version = '1.4'
 module_pathname = '$libdir/aqo'
 relocatable = false

expected/forced_stat_collection.out

@@ -36,8 +36,8 @@ SELECT count(*) FROM person WHERE age<18 AND passport IS NOT NULL;
 (1 row)
 
 SELECT * FROM aqo_data;
- fspace_hash | fsspace_hash | nfeatures | features | targets | oids 
--------------+--------------+-----------+----------+---------+------
+ fspace_hash | fsspace_hash | nfeatures | features | targets | oids | reliability 
+-------------+--------------+-----------+----------+---------+------+-------------
 (0 rows)
 
 CREATE OR REPLACE FUNCTION round_array (double precision[])

learn_cache.c

@@ -33,7 +33,8 @@ typedef struct
 	int			nrows;
 	int			ncols;
 	double	   *matrix[aqo_K];
-	double	   *targets;
+	double		targets[aqo_K];
+	double		rfactors[aqo_K];
 	List	   *relids;
 } htab_entry;
 
@@ -76,7 +77,6 @@ lc_update_fss(uint64 fs, int fss, OkNNrdata *data, List *relids)
 		/* Clear previous version of the cached data. */
 		for (i = 0; i < entry->nrows; ++i)
 			pfree(entry->matrix[i]);
-		pfree(entry->targets);
 		list_free(entry->relids);
 	}
 
@@ -87,8 +87,9 @@ lc_update_fss(uint64 fs, int fss, OkNNrdata *data, List *relids)
 		entry->matrix[i] = palloc(sizeof(double) * data->cols);
 		memcpy(entry->matrix[i], data->matrix[i], sizeof(double) * data->cols);
 	}
-	entry->targets = palloc(sizeof(double) * data->rows);
+
 	memcpy(entry->targets, data->targets, sizeof(double) * data->rows);
+	memcpy(entry->rfactors, data->rfactors, sizeof(double) * data->rows);
 	entry->relids = list_copy(relids);
 
 	MemoryContextSwitchTo(memctx);
@@ -137,6 +138,7 @@ lc_load_fss(uint64 fs, int fss, OkNNrdata *data, List **relids)
 	for (i = 0; i < entry->nrows; ++i)
 		memcpy(data->matrix[i], entry->matrix[i], sizeof(double) * data->cols);
 	memcpy(data->targets, entry->targets, sizeof(double) * entry->nrows);
+	memcpy(data->rfactors, entry->rfactors, sizeof(double) * entry->nrows);
 	if (relids)
 		*relids = list_copy(entry->relids);
 	return true;
@@ -165,7 +167,7 @@ lc_remove_fss(uint64 fs, int fss)
 
 	for (i = 0; i < entry->nrows; ++i)
 		pfree(entry->matrix[i]);
-	pfree(entry->targets);
+
 	hash_search(fss_htab, &key, HASH_REMOVE, NULL);
 }
 

machine_learning.c

@@ -156,8 +156,7 @@ OkNNr_predict(OkNNrdata *data, double *features)
  * partially executed statement.
  */
 int
-OkNNr_learn(OkNNrdata *data,
-			double *features, double target, double rfactor)
+OkNNr_learn(OkNNrdata *data, double *features, double target, double rfactor)
 {
 	double	distances[aqo_K];
 	int		i;
@@ -191,10 +190,10 @@ OkNNr_learn(OkNNrdata *data,
 
 	if (data->rows < aqo_K)
 	{
-		/* We can't reached limit of stored neighbors */
+		/* We don't reach a limit of stored neighbors */
 
 		/*
-		 * Add new line into the matrix. We can do this because matrix_rows
+		 * Add new line into the matrix. We can do this because data->rows
 		 * is not the boundary of matrix. Matrix has aqo_K free lines
 		 */
 		for (j = 0; j < data->cols; ++j)
@@ -206,7 +205,7 @@ OkNNr_learn(OkNNrdata *data,
 	}
 	else
 	{
-		double	*feature;
+		double *feature;
 		double	avg_target = 0;
 		double	tc_coef; /* Target correction coefficient */
 		double	fc_coef; /* Feature correction coefficient */

storage.c

@@ -27,6 +27,7 @@
 #include "learn_cache.h"
 
 
+#define AQO_DATA_COLUMNS	(7)
 HTAB *deactivated_queries = NULL;
 
 static ArrayType *form_matrix(double **matrix, int nrows, int ncols);
@@ -402,8 +403,8 @@ load_fss(uint64 fs, int fss, OkNNrdata *data, List **relids)
 	bool		find_ok = false;
 	IndexScanDesc scan;
 	ScanKeyData	key[2];
-	Datum		values[6];
-	bool		isnull[6];
+	Datum		values[AQO_DATA_COLUMNS];
+	bool		isnull[AQO_DATA_COLUMNS];
 	bool		success = true;
 
 	if (!open_aqo_relation("public", "aqo_data",
@@ -434,6 +435,7 @@ load_fss(uint64 fs, int fss, OkNNrdata *data, List **relids)
 				deform_matrix(values[3], data->matrix);
 
 			deform_vector(values[4], data->targets, &(data->rows));
+			deform_vector(values[6], data->rfactors, &(data->rows));
 
 			if (relids != NULL)
 				*relids = deform_oids_vector(values[5]);
@@ -487,9 +489,9 @@ update_fss(uint64 fhash, int fsshash, OkNNrdata *data, List *relids)
 	TupleDesc	tupDesc;
 	HeapTuple	tuple,
 				nw_tuple;
-	Datum		values[6];
-	bool		isnull[6] = { false, false, false, false, false, false };
-	bool		replace[6] = { false, false, false, true, true, false };
+	Datum		values[AQO_DATA_COLUMNS];
+	bool		isnull[AQO_DATA_COLUMNS];
+	bool		replace[AQO_DATA_COLUMNS] = { false, false, false, true, true, false, true };
 	bool		shouldFree;
 	bool		find_ok = false;
 	bool		update_indexes;
@@ -506,6 +508,7 @@ update_fss(uint64 fhash, int fsshash, OkNNrdata *data, List *relids)
 						   RowExclusiveLock, &hrel, &irel))
 		return false;
 
+	memset(isnull, 0, sizeof(bool) * AQO_DATA_COLUMNS);
 	tupDesc = RelationGetDescr(hrel);
 	InitDirtySnapshot(snap);
 	scan = index_beginscan(hrel, irel, &snap, 2, 0);
@@ -535,6 +538,7 @@ update_fss(uint64 fhash, int fsshash, OkNNrdata *data, List *relids)
 		values[5] = PointerGetDatum(form_oids_vector(relids));
 		if ((void *) values[5] == NULL)
 			isnull[5] = true;
+		values[6] = PointerGetDatum(form_vector(data->rfactors, data->rows));
 		tuple = heap_form_tuple(tupDesc, values, isnull);
 
 		/*
@@ -558,8 +562,8 @@ update_fss(uint64 fhash, int fsshash, OkNNrdata *data, List *relids)
 			isnull[3] = true;
 
 		values[4] = PointerGetDatum(form_vector(data->targets, data->rows));
-		nw_tuple = heap_modify_tuple(tuple, tupDesc,
-									 values, isnull, replace);
+		values[6] = PointerGetDatum(form_vector(data->rfactors, data->rows));
+		nw_tuple = heap_modify_tuple(tuple, tupDesc, values, isnull, replace);
 		if (my_simple_heap_update(hrel, &(nw_tuple->t_self), nw_tuple,
 															&update_indexes))
 		{