XRP Ledger APIs generally use strings, rather than native JSON numbers, to represent numeric amounts of currency for both XRP and issued currencies. This protects against a loss of precision when using JSON parsers, which may automatically try to represent all JSON numbers in a floating-point format. Within the String value, the numbers are serialized in the same way as native JSON numbers:
- May contain
.as a decimal point. For example, ½ is represented as
0.5. (American style, not European)
- May contain
eto indicate being raised to a power of 10 (scientific notation). For example,
1.2E5is equivalent to 1.2×105, or
- No comma (
,) characters are used.
XRP has the same precision as a 64-bit unsigned integer where each unit is equivalent to 0.000001 XRP. Its properties are:
- Minimum value:
- Maximum value:
"100000000000000000"(1017) drops of XRP
- Precise to the nearest
"1"drop of XRP
Issued Currency Precision
Issued currencies in the XRP Ledger are represented with a custom format with the following precision:
- Minimum nonzero absolute value:
- Maximum value:
- Minimum value:
- 15 decimal digits of precision
Issued Currency Math
rippled represents numbers for issued currencies in a custom number format. This format can store a wide variety of assets, including those typically measured in very small or very large denominations. Unlike typical floating-point representations of non-whole numbers, this format uses integer math for all calculations, so it always maintains 15 decimal digits of precision. Unlike "arbitrary precision" number formats, the custom format can always be stored in a fixed size of 64 bits.
The internal format consists of three parts: a sign bit, significant digits, and an exponent. (It uses them in the same way as scientific notation.) The sign bit indicates whether the amount is positive or negative. The significant digits are represented using an integer in the range
9999999999999999 (inclusive), except for the special case of the value 0, whose significant digits use the value
0. The exponent indicates the scale (what power of 10 the significant digits should be multiplied by) in the range -96 to +80 (inclusive). Before recording any amount,
rippled "canonicalizes" the value so that the significant digits and exponent are within the expected range. For example, the canonical representation of 1 unit of currency is
1000000000000000e-15. The internal calculations generally use integer math so that numbers are always precise within 15 digits. Multiplication and division have adjustments to compensate for over-rounding in the least significant digits.
When transmitting non-XRP amounts across the network or recording them in ledgers, the amounts are joined into a 64-bit format. The most significant bit indicates whether the amount is XRP or issued currency. (The value
1 indicates a non-XRP amount.) The next bit is the sign bit, 1 for positive or 0 for negative. (Caution: This is the opposite of how sign bits work in most other numeric representations!) The next 8 bits are the exponent, and the significant digits occupy the remaining 54 bits.
All non-XRP currencies in the XRP Ledger have a 160-bit currency code. The
rippled APIs map 3-character ASCII strings (case-sensitive) to 160-bit currency codes using a standard mapping. The currency code
XRP is disallowed for issued currencies. Currencies with the same code can ripple across connected trustlines. Currency codes have no other behavior built into the XRP Ledger.
Standard Currency Codes
The standard currency mapping allocates the bits as follows:
- The first 8 bits must be
- The next 96 bits are reserved, and should be all
- The next 24 bits represent 3 characters of ASCII.
Ripple recommends using ISO 4217 codes, or popular pseudo-ISO 4217 codes such as "BTC". However, any combination of the following characters is permitted: all uppercase and lowercase letters, digits, as well as the symbols
|. The currency code
XRP(all-uppercase) is reserved for XRP and cannot be used by issued currencies.
- The next 8 bits indicate the currency version. If the same currency is reissued with a different value, you can increment this value to keep the currencies separate.
- The next 24 bits are reserved and should be all
Nonstandard Currency Codes
You can also issue currency of other types by using a 160-bit (40-character) hexadecimal string such as
015841551A748AD2C1F76FF6ECB0CCCD00000000 as the currency code. To prevent this from being treated as a different currency code type, the first 8 bits MUST NOT be
Deprecated: Some previous versions of ripple-lib supported an "interest-bearing" or "demurraging" currency code type. These currencies have the first 8 bits
0x01. Demurraging / interest-bearing currencies are no longer supported, but you may encounter them in ledger data. For more information, see Demurrage.