BytesHash
Contains a ptr to the string slice an the precomputed hash of that string. During rehashes, we will rehash the hash instead of the string, that makes rehashing cheap and allows cache coherent small hash tables.
IdHasher
IdxHash
Contains an idx of a row in a DataFrame and the precomputed hash of that row. That hash still needs to be used to create another hash to be able to resize hashmaps without accidental quadratic behavior. So do not use an Identity function!

Constants

Traits

_boost_hash_combine
_df_rows_to_hashes_threaded_vertical
_hash_binary_array
integer_hash
populate_multiple_key_hashmap
Populate a multiple key hashmap with row indexes. Instead of the keys (which could be very large), the row indexes are stored. To check if a row is equal the original DataFrame is also passed as ref. When a hash collision occurs the indexes are ptrs to the rows and the rows are compared on equality.