Module adaptive 

Adaptive encoding selection — samples data to choose the optimal encoding.

AdaptiveSelector examines the first N values (default: 1024) of a column to estimate which encoder will achieve the best compression ratio, then encodes the full column with that encoder.

Decision matrix

Pattern	Selected encoding
All values identical	RLE
Periodic floats	Chimp128
Regular intervals (i64)	Delta-of-delta
Narrow-range integers	Frame-of-Reference (FOR)
Slowly varying floats	Chimp128 / Chimp
Random floats	Gorilla → Plain fallback
Low-cardinality strings	Dictionary
High-cardinality strings	Plain
Booleans	Bitmap (always)

Structs

AdaptiveSelector

Adaptive encoding selector.

Enums

FloatPattern

Result of analyzing a float sample.

IntegerPattern

Result of analyzing an integer sample.

StringPattern

Result of analyzing a string sample.