Module function 

iceberg_hash, the hash allowed on the column side of a partition rule.

The hash decides which region a row is written to and is persisted in partition metadata, so its result for a given input must never change across versions. A different algorithm needs a new type.

Rule simplification and coverage checks treat the hash as an independent dimension, which holds because it is defined for every value of the accepted column types.

Structs

IcebergHash

iceberg_hash(column): the 32-bit Murmur3 hash defined by the Iceberg table spec (appendix B) over a string or integer column, as an INT.

IcebergHashUdf 🔒

DataFusion form of IcebergHash, so storage engines can evaluate a region’s partition expression as a filter.

Functions

hash_long 🔒

iceberg_hash_array 🔒

is_signed_integer 🔒

murmur3_32 🔒

MurmurHash3 x86 32-bit with seed 0, returned as a signed integer like Iceberg.