How can I speed up the following sql query? [duplicate] - sql

Question

This question already exists:
Closed 11 years ago.
Possible Duplicate:
How can I speed up the following update query?
I would like to run the following query in an acceptable time (e.g. max 15 mins) on a modern desktop computer running postgresql 8.4:
UPDATE cap_options_rule_items
SET cap_option_id = O.id
FROM cap_options_rule_items I
JOIN cap_options_rules R
ON R.id = I.cap_options_rule_id
JOIN cap_options O
ON R.cap_engine_id = O.cap_engine_id
AND O.code = I.cap_option_code;
I would like to know if there are obvious mistakes I'm doing in the query on with the choice of indexes.
The tables in the query have the following number of records:
cap_options_rule_item: 2208705
cap_options_rule: 430268
cap_options: 1628188
And the following schema (including indexes)
-- Table: cap_options_rule_items
CREATE TABLE cap_options_rule_items
(
id serial NOT NULL,
cap_options_rule_id integer,
cap_option_code integer,
"primary" boolean,
cap_option_id integer,
CONSTRAINT cap_options_rule_items_pkey PRIMARY KEY (id)
)
WITH (
OIDS=FALSE
);
-- Index: index_cap_options_rule_items_on_cap_option_id
CREATE INDEX index_cap_options_rule_items_on_cap_option_id
ON cap_options_rule_items
USING btree (cap_option_code);
-- Index: index_cap_options_rule_items_on_cap_option_rule_id
CREATE INDEX index_cap_options_rule_items_on_cap_option_rule_id
ON cap_options_rule_items
USING btree (cap_options_rule_id);
-- Table: cap_options_rules
CREATE TABLE cap_options_rules
(
id serial NOT NULL,
rule_type character varying(255),
cap_engine_id integer,
CONSTRAINT cap_options_rules_pkey PRIMARY KEY (id)
) WITH ( OIDS=FALSE
);
-- Index: index_cap_options_rules_on_cap_engine_id
CREATE INDEX index_cap_options_rules_on_cap_engine_id
ON cap_options_rules
USING btree (cap_engine_id);
-- Table: cap_options
CREATE TABLE cap_options
( id serial NOT NULL,
description character varying(255),
cap_engine_id integer,
cap_option_category_id integer,
basic_price numeric,
vat numeric,
default_option boolean,
created_at timestamp without time zone,
updated_at timestamp without time zone,
code integer,
CONSTRAINT cap_options_pkey PRIMARY KEY (id)
) WITH ( OIDS=FALSE
);
-- Index: index_code_and_cap_engine_id_on_cap_options
CREATE INDEX index_code_and_cap_engine_id_on_cap_options
ON cap_options
USING btree (code, cap_engine_id);
Thank you!

Answer 1

Your query is slow because you are updating all the rows in cap_options_rule_items.
I think you really want something like this:
UPDATE cap_options_rule_items I
SET cap_option_id = O.id
FROM cap_options_rules R
join cap_options O on R.cap_engine_id = O.cap_engine_id
WHERE I.cap_options_rule_id = R.id
and I.cap_option_code = O.code;