DOUBLE PRECISION

PostgreSQL's DOUBLE PRECISION type (also called FLOAT8 or FLOAT) is an inexact, variable-precision numeric type implementing the IEEE 754 double-precision standard. It is best suited for scientific measurements and calculations where approximate results are acceptable.

Key characteristics

• Storage: 8 bytes
• Range: approximately 1E-307 to 1E+308
• Precision: at least 15 significant decimal digits
• Synonyms: DOUBLE PRECISION, FLOAT8, FLOAT

Basic usage: storing temperature readings

CREATE TABLE temperatures (
  id          SERIAL PRIMARY KEY,
  location    TEXT NOT NULL,
  temperature DOUBLE PRECISION
);

INSERT INTO temperatures (location, temperature)
VALUES
  ('Lab Room 1',    23.5),
  ('Server Room 1', 21.8),
  ('Server Room 2', 24.3)
RETURNING *;

id |   location    | temperature
----+---------------+-------------
 1 | Lab Room 1    |        23.5
 2 | Server Room 1 |        21.8
 3 | Server Room 2 |        24.3

SELECT AVG(temperature) FROM temperatures;

  avg
------
 23.2

The inexact nature of DOUBLE PRECISION

Because DOUBLE PRECISION uses binary floating-point, some decimal fractions cannot be stored exactly:

CREATE TABLE t (c double precision);

INSERT INTO t(c) VALUES (0.1), (0.1), (0.1);

SELECT SUM(c) FROM t;

        sum
---------------------
 0.30000000000000004

This is expected IEEE 754 behavior. For financial data, use NUMERIC instead.

Numbers outside the valid range cause an error:

INSERT INTO t(c) VALUES (1E-400);
-- ERROR: "0.000...0001" is out of range for type double precision

Production tips

• Never use DOUBLE PRECISION for monetary values — use NUMERIC to avoid floating-point rounding errors in financial calculations.
• Avoid equality comparisons with DOUBLE PRECISION values (WHERE temp = 23.5) — use range comparisons instead (WHERE temp BETWEEN 23.4 AND 23.6).
• Use DOUBLE PRECISION for scientific measurements, GPS coordinates, and machine learning feature values where approximate precision is acceptable.
• Arithmetic on DOUBLE PRECISION is significantly faster than NUMERIC, making it appropriate for analytical workloads on large datasets.

Reference: PostgreSQL documentation — Floating-Point Types.