PostgreSQL – Using Window Functions to Replace Multiple Subqueries

postgresqlsubquery

I have two tables:

CREATE TABLE fund_data
(
  id serial NOT NULL,
  fund_entries_id integer NOT NULL,
  fund_val numeric(25,6) NOT NULL,
  bbg_pulls_id integer NOT NULL,
  CONSTRAINT fund_data_pkey PRIMARY KEY (id),
  CONSTRAINT fund_data_bbg_pulls_id_fkey FOREIGN KEY (bbg_pulls_id)
      REFERENCES bbg_pulls (id) MATCH SIMPLE
      ON UPDATE NO ACTION ON DELETE NO ACTION,
  CONSTRAINT fund_data_fund_entries_id_fkey FOREIGN KEY (fund_entries_id)
      REFERENCES fund_entries (id) MATCH SIMPLE
      ON UPDATE NO ACTION ON DELETE NO ACTION,
  CONSTRAINT fund_data_fund_entries_id_bbg_pulls_id_key UNIQUE (fund_entries_id, bbg_pulls_id)
)

And:

CREATE TABLE fund_entries
(
  id serial NOT NULL,
  fiscal_prd_end_date date NOT NULL,
  company_id integer NOT NULL,
  ern_dt_id integer NOT NULL,
  CONSTRAINT fund_entries_pkey PRIMARY KEY (id),
  CONSTRAINT fund_entries_company_id_fkey FOREIGN KEY (company_id)
      REFERENCES company (id) MATCH SIMPLE
      ON UPDATE NO ACTION ON DELETE NO ACTION,
  CONSTRAINT fund_entries_ern_dt_id_fkey11 FOREIGN KEY (ern_dt_id)
      REFERENCES ern_dt (id) MATCH SIMPLE
      ON UPDATE NO ACTION ON DELETE NO ACTION,
  CONSTRAINT fund_entries_company_id_fiscal_prd_end_date_key UNIQUE (company_id, fiscal_prd_end_date),
  CONSTRAINT fund_entries_ern_dt_id_key UNIQUE (ern_dt_id)
)

To get the most recent value from fund_data for a given company_id and bbg_pulls_id, I currently do two correlated sub-queries:

SELECT fd.fund_val
FROM fund_data fd
    ,(
        SELECT fe.id
        FROM fund_entries fe
            ,(
                SELECT max(fe.fiscal_prd_end_date)
                FROM fund_entries fe
                WHERE fe.company_id = 858
                ) md
        WHERE fe.fiscal_prd_end_date = md.max
            AND fe.company_id = 858
        ) mi
WHERE fd.bbg_pulls_id = 20
    AND fd.fund_entries_id = mi.id

Question: The query works find, but can I use windowing to avoid the two sub-queries? I've tried several times and I keep getting multiple rows.

EDIT: A description of the query in words: Find the max date for a given company_id, so I can find the id for that record. Then use that id to get the most recent value associated with that fk id (and max date) for a given bbg_pulls_id and company. Hopefully that helps.

Best Answer

If I understand correctly:

SELECT DISTINCT
     last_value(fd.fund_val) OVER (ORDER BY fe.fiscal_prd_end_date)
FROM
     fund_data AS fd
     JOIN fund_entries AS fe ON (fd.fund_entries_id = fe.id)
WHERE
     fd.bbg_pulls_id = 20
     AND fe.company_id = 858

UPDATED:

Or just:

SELECT
     fd.fund_val
FROM
     fund_data AS fd
     JOIN fund_entries AS fe ON (fd.fund_entries_id = fe.id)
WHERE
     fd.bbg_pulls_id = 20
     AND fe.company_id = 858
ORDER BY fe.fiscal_prd_end_date DESC
LIMIT 1

Related Solutions

Postgresql – Postgres multiple joins slow query, how to store default child record

You write:

Each customer can have multiple sites, but only one should be displayed in this list.

Yet, your query retrieves all rows. That would be a point to optimize. But you also do not define which site is to be picked.

Either way, it does not matter much here. Your EXPLAIN shows only 5026 rows for the site scan (5018 for the customer scan). So hardly any customer actually has more than one site. Did you ANALYZE your tables before running EXPLAIN?

From the numbers I see in your EXPLAIN, indexes will give you nothing for this query. Sequential table scans will be the fastest possible way. Half a second is rather slow for 5000 rows, though. Maybe your database needs some general performance tuning?

Maybe the query itself is faster, but "half a second" includes network transfer? EXPLAIN ANALYZE would tell us more.

If this query is your bottleneck, I would suggest you implement a materialized view.

After you provided more information I find that my diagnosis pretty much holds.

The query itself needs 27 ms. Not much of a problem there. "Half a second" was the kind of misunderstanding I had suspected. The slow part is the network transfer (plus ssh encoding / decoding, possibly rendering). You should only retrieve 100 rows, that would solve most of it, even if it means to execute the whole query every time.

If you go the route with a materialized view like I proposed you could add a serial number without gaps to the table plus index on it - by adding a column row_number() OVER (<your sort citeria here>) AS mv_id.

Then you can query:

SELECT *
FROM   materialized_view
WHERE  mv_id >= 2700
AND    mv_id <  2800;

This will perform very fast. LIMIT / OFFSET cannot compete, that needs to compute the whole table before it can sort and pick 100 rows.

pgAdmin timing

When you execute a query from the query tool, the message pane shows something like:

Total query runtime: 62 ms.

And the status line shows the same time. I quote pgAdmin help about that:

The status line will show how long the last query took to complete. If a dataset was returned, not only the elapsed time for server execution is displayed, but also the time to retrieve the data from the server to the Data Output page.

If you want to see the time on the server you need to use SQL EXPLAIN ANALYZE or the built in Shift + F7keyboard shortcut or Query -> Explain analyze. Then, at the bottom of the explain output you get something like this:

Total runtime: 0.269 ms

Postgresql – What happens to the index of a primary key after a DROP CONSTRAINT

Disclaimer

This is experimental and only tested rudimentarily. Proceed at your own risk. I would not use it myself and just drop / recreate constraints with standard DDL commands. If you break entries in the catalog tables you could easily mess up your database.

For all I know, there are only two differences between a PRIMARY KEY and a UNIQUE constraint in the catalog tables (the index itself is identical):

pg_index.indisprimary:
For PRIMARY KEY constraint ... TRUE
For UNIQUE constraint ... FALSE

pg_constraint.contype:
PRIMARY KEY constraint ... 'p'
UNIQUE constraint ... 'u'

You could convert constraint and index in place, from PRIMARY KEY constraint to UNIQUE constraint, my_idx being the (optionally schema-qualified) index name:

UPDATE pg_index SET indisprimary = FALSE WHERE indexrelid = 'my_idx'::regclass
UPDATE pg_constraint SET contype = 'u' WHERE conindid = 'my_idx'::regclass;

Or upgrade from UNIQUE to PRIMARY KEY:

UPDATE pg_index SET indisprimary = TRUE WHERE indexrelid = 'my_idx'::regclass;
UPDATE pg_constraint SET contype = 'p' WHERE conindid = 'my_idx'::regclass;