COVAR_SAMP

Overview

The COVAR_SAMP() aggregate function calculates the sample covariance between two sets of number pairs. This function measures how changes in one variable relate linearly to changes in another variable within a sample dataset.

Syntax

The syntax for this function is as follows:

COVAR_SAMP(y, x)

Parameters

• y: variable being predicted
• x: variable used for prediction

Example

For the needs of this section, we’re going to use a simplified version of the film table from the Pagila database, containing only the title, length and rating columns. The complete schema for the film table can be found on the Pagila database website.

DROP TABLE IF EXISTS film;
CREATE TABLE film (
  title text NOT NULL,
  length int,
  rating int
);
INSERT INTO film(title, length, rating) VALUES
  ('ATTRACTION NEWTON', 83, 5),
  ('CHRISTMAS MOONSHINE', 150, 7),
  ('DANGEROUS UPTOWN', 121, 4),
  ('KILL BROTHERHOOD', 54, 3),
  ('HALLOWEEN NUTS', 47, 5),
  ('HOURS RAGE', 122, 7),
  ('PIANIST OUTFIELD', 136, 7),
  ('PICKUP DRIVING', 77, 3),
  ('INDEPENDENCE HOTEL', 157, 7),
  ('PRIVATE DROP', 106, 4),
  ('SAINTS BRIDE', 125, 3),
  ('FOREVER CANDIDATE', 131, 7),
  ('MILLION ACE', 142, 5),
  ('SLEEPY JAPANESE', 137, 4),
  ('WRATH MILE', 176, 7),
  ('YOUTH KICK', 179, 7),
  ('CLOCKWORK PARADISE', 143, 5);

The query below query uses the COVAR_SAMP() function to calculate the sample covariance between film length and rating where rating is greater than or equal to 4:

SELECT
    COVAR_SAMP(length, rating) AS SampleCovariance
FROM film
WHERE rating >= 4;

By running the above query will get the following output:

  samplecovariance  
--------------------
 23.087912087912066
(1 row)