name: type-f64
kind: type
description: IEEE-754 double-precision float; ~15 significant digits.
type: reference
applies_to: inspect, predict
Bytes
Fixed-width: 8 bytes per record, little-endian IEEE-754 binary64. Not bit-packed. Stride contributes 8 bytes; reader uses math.Float64frombits(binary.LittleEndian.Uint64(...)).
Range
±1.8e308 with ~15 decimal significant digits. The analytics default for any continuous numeric — scores, probabilities, summary statistics, model outputs. Sufficient for most scientific computing; reach for decimal128 only when exact base-10 semantics matter (money, regulated metrics).
Null
Orthogonal. Nullable: true participates in the per-record null bitmap. NaN is the IEEE NaN value, not null; 0.0 is the float zero, not null. IsNumeric() and IsNumericForAnalytics() are both true; aggregators apply null-skip semantics.
Dictionary
Absent. Floats are never dictionary-encoded.
See
- Skill:
cohort-schema-design(Field-type matrix, Selection heuristics). - Cross-link:
type-decimal128for exact base-10 arithmetic;financial-cohorts.