Google BigQuery: The Definitive Guide

Chapter 3. Data Types, Functions, and Operators

In the bike rental queries in the previous chapters, when we divided the trip duration by 60, we were able to do so because trip duration was a numeric type. Trying to divide the gender by 60 would not have worked because gender is a string. The functions and operations you have at your disposal might be restricted based on the type of data to which you are applying them.

BigQuery supports several data types to store numeric, string, time, geographic, structured, and semi-structured data:

INT64

This is the only integer type. It can represent numbers ranging from approximately 10^–19 to 10¹⁹. For real-valued numbers, use FLOAT64, and for Booleans, use BOOL.

NUMERIC

NUMERIC offers 38 digits of precision and 9 decimal digits of scale and is suitable for exact calculations, such as in finance.

STRING

This is a first-class type and represents variable-length sequences of Unicode characters. BYTES are variable-length sequences of characters (not Unicode).

TIMESTAMP

This represents an absolute point in time.

DATETIME

This represents a calendar date and time. DATE and TIME are also available separately.

GEOGRAPHY

GEOGRAPHY represents points, lines, and polygons on the surface of the Earth.

STRUCT and ARRAY

See the description for each of these in Chapter 2.

Numeric Types and Functions

As just stated, there is only one integer type (INT64) and only one floating-point type (FLOAT64). Both of these types support the typical arithmetic operations ...