Enum DecimalView 

pub enum DecimalView<'a> {
    String {
        value: &'a str,
    },
    Scaled {
        scale: u8,
        value: Cow<'a, [u8]>,
    },
}

A decimal value backed by either a string representation or a scaled mantissa.

Decimal values can be serialized in two formats:

Text Format

The decimal is written as a string representation followed by a 'd' suffix.

Example: "123.45d" or "1.5e-3d"

Binary Format

A more compact binary encoding consisting of:

1. Binary format marker: '=' (0x3D)
2. Type identifier: DECIMAL_BINARY_FORMAT_TYPE byte
3. Scale: 1 byte (0-76 inclusive) - number of decimal places
4. Length: 1 byte - number of bytes in the unscaled value
5. Unscaled value: variable-length byte array in two’s complement format, big-endian

Example: For decimal 123.45 with scale 2:

Unscaled value: 12345
Binary representation:
  = [23] [2] [2] [0x30] [0x39]
  │  │    │   │  └───────────┘
  │  │    │   │        └─ Mantissa bytes (12345 in big-endian)
  │  │    │   └─ Length: 2 bytes
  │  │    └─ Scale: 2
  │  └─ Type: DECIMAL_BINARY_FORMAT_TYPE (23)
  └─ Binary marker: '='

Binary Format Notes

• The unscaled value must be encoded in two’s complement big-endian format
• Maximum scale is 76
• Length byte indicates how many bytes follow for the unscaled value

Variants

String

Fields

value: &'a str

Scaled

Fields

scale: u8

value: Cow<'a, [u8]>

Implementations

impl<'a> DecimalView<'a>

pub fn try_new_scaled<T>(scale: u32, value: T) -> Result<Self>

where T: Into<Cow<'a, [u8]>>,

Creates a DecimalView::Scaled from a mantissa buffer and scale.

Validates that:

• scale does not exceed the QuestDB maximum of 76 decimal places.
• The mantissa fits into at most 32 bytes (ILP binary limit).

Returns an error::ErrorCode::InvalidDecimal error if either constraint is violated.

pub fn try_new_string(value: &'a str) -> Result<Self>

Creates a DecimalView::String from a textual decimal representation.

Thousand separators (commas) are not allowed and the decimal point must be a dot (.).

Performs lightweight validation and rejects values containing ILP-reserved characters. Accepts plain decimals, optional +/- prefixes, NaN, Infinity, and scientific notation (e/E).

Returns error::ErrorCode::InvalidDecimal if disallowed characters are encountered.

Trait Implementations

impl<'a> Debug for DecimalView<'a>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'a> TryInto<DecimalView<'a>> for &'a BigDecimal

Available on crate feature bigdecimal only.

type Error = Error

The type returned in the event of a conversion error.

fn try_into(self) -> Result<DecimalView<'a>>

Performs the conversion.

impl<'a> TryInto<DecimalView<'a>> for &'a Decimal

Available on crate feature rust_decimal only.

type Error = Error

The type returned in the event of a conversion error.

fn try_into(self) -> Result<DecimalView<'a>>

Performs the conversion.

impl<'a> TryInto<DecimalView<'a>> for &'a str

Implementation for string slices containing decimal representations.

This implementation uses the text format.

Format

The string is validated and written as-is, followed by the ‘d’ suffix. Thousand separators (commas) are not allowed and the decimal point must be a dot (.).

Validation

The implementation performs partial validation only:

• Rejects non-numerical characters (not -/+, 0-9, ., Infinity, NaN, e/E)
• Does NOT validate the actual decimal syntax (e.g., “e2e” would pass)

This is intentional: full parsing would add overhead. The QuestDB server performs complete validation and will reject malformed decimals.

Examples

• "123.45" → "123.45d"
• "1.5e-3" → "1.5e-3d"
• "-0.001" → "-0.001d"

Errors

Returns [Error] with ErrorCode::InvalidDecimal if the string contains non-numerical characters.

type Error = Error

The type returned in the event of a conversion error.

fn try_into(self) -> Result<DecimalView<'a>>

Performs the conversion.