A representation of a decimal value, such as 2.5. Clients may convert values into language-native decimal formats, such as Java's [BigDecimal][] or Python's [decimal.Decimal][].

[BigDecimal]: [decimal.Decimal]:

JSON representation
  "value": string


The decimal value, as a string.

The string representation consists of an optional sign, + (U+002B) or - (U+002D), followed by a sequence of zero or more decimal digits ("the integer"), optionally followed by a fraction, optionally followed by an exponent. An empty string should be interpreted as 0.

The fraction consists of a decimal point followed by zero or more decimal digits. The string must contain at least one digit in either the integer or the fraction. The number formed by the sign, the integer and the fraction is referred to as the significand.

The exponent consists of the character e (U+0065) or E (U+0045) followed by one or more decimal digits.

Services should normalize decimal values before storing them by:

  • Removing an explicitly-provided + sign (+2.5 -> 2.5).
  • Replacing a zero-length integer value with 0 (.5 -> 0.5).
  • Coercing the exponent character to upper-case, with explicit sign (2.5e8 -> 2.5E+8).
  • Removing an explicitly-provided zero exponent (2.5E0 -> 2.5).

Services may perform additional normalization based on its own needs and the internal decimal implementation selected, such as shifting the decimal point and exponent value together (example: 2.5E-1 <-> 0.25). Additionally, services may preserve trailing zeroes in the fraction to indicate increased precision, but are not required to do so.

Note that only the . character is supported to divide the integer and the fraction; , should not be supported regardless of locale. Additionally, thousand separators should not be supported. If a service does support them, values must be normalized.

The ENBF grammar is:

DecimalString =
  '' | [Sign] Significand [Exponent];

Sign = '+' | '-';

Significand =
  Digits ['.'] [Digits] | [Digits] '.' Digits;

Exponent = ('e' | 'E') [Sign] Digits;

Digits = { '0' | '1' | '2' | '3' | '4' | '5' | '6' | '7' | '8' | '9' };

Services should clearly document the range of supported values, the maximum supported precision (total number of digits), and, if applicable, the scale (number of digits after the decimal point), as well as how it behaves when receiving out-of-bounds values.

Services may choose to accept values passed as input even when the value has a higher precision or scale than the service supports, and should round the value to fit the supported scale. Alternatively, the service may error with 400 Bad Request (INVALID_ARGUMENT in gRPC) if precision would be lost.

Services should error with 400 Bad Request (INVALID_ARGUMENT in gRPC) if the service receives a value outside of the supported range.