Package it.unimi.dsi.sux4j.mph

Class CHDMinimalPerfectHashFunction<T>

java.lang.Object
it.unimi.dsi.fastutil.objects.AbstractObject2LongFunction<K>
it.unimi.dsi.sux4j.mph.AbstractHashFunction<T>
it.unimi.dsi.sux4j.mph.CHDMinimalPerfectHashFunction<T>
All Implemented Interfaces:
Function<T,Long>, Object2LongFunction<T>, Size64, Serializable, Function<T,Long>, ToLongFunction<T>
public class CHDMinimalPerfectHashFunction<T> extends AbstractHashFunction<T> implements Serializable
A minimal perfect hash function implemented using the “hash, displace and compress” technique.

Given a list of keys without duplicates, the builder of this class finds a minimal perfect hash function for the list. Subsequent calls to the getLong(Object) method will return a distinct number for each key in the list. For keys out of the list, the resulting number is not specified. In some (rare) cases it might be possible to establish that a key was not in the original list, and in that case -1 will be returned; by signing the function (see below), you can guarantee with a prescribed probability that -1 will be returned on keys not in the original list. The class can then be saved by serialisation and reused later.

This class uses a chunked hash store to provide highly scalable construction. Note that at construction time you can pass a it.unimi.dsi.sux4j.io.ChunkedHashStore containing the keys (associated with any value); however, if the store is rebuilt because of a DuplicateException it will be rebuilt associating with each key its ordinal position.

The memory requirements for the algorithm we use are ≈2 bits per key for load factor equal to one and λ = 5. Thus, this class can use ≈10% less memory than a GOVMinimalPerfectHashFunction.

However, its construction time is an order of magnitude larger, and query time is about 50% slower. Different tradeoffs between construction time, query time and space can be obtained by tweaking the load factor and the parameter λ (see the paper below for their exact meaning).

For convenience, this class provides a main method that reads from standard input a (possibly gzip'd) sequence of newline-separated strings, and writes a serialised minimal perfect hash function for the given list.

Signing

Optionally, it is possible to sign the minimal perfect hash function. A w-bit signature will be associated with each key, so that getLong(Object) will return -1 on strings that are not in the original key set. As usual, false positives are possible with probability 2-w.

How it Works

The technique used is described by Djamal Belazzougui, Fabiano C. Botelho and Martin Dietzfelbinger in “Hash, displace and compress”, Algorithms - ESA 2009, LNCS 5757, pages 682−693, 2009. However, with respect to the algorithm described in the paper, this implementation is much more scalable, as it uses a ChunkedHashStore to split the generation of large key sets into generation of smaller functions for each chunk (of size approximately 216).

Since:
3.2.0
Author:
Sebastiano Vigna
See Also:

  • Field Details

    • serialVersionUID

      public static final long serialVersionUID
      See Also:

    • LOG2_CHUNK_SIZE

      public static final int LOG2_CHUNK_SIZE
      The logarithm of the desired chunk size.
      See Also:

    • n

      protected final long n
      The number of keys.

    • globalSeed

      protected final long globalSeed
      The seed used to generate the initial hash triple.

    • transform

      protected final TransformationStrategy<? super T> transform
      The transformation strategy.

    • coefficients

      protected final EliasFanoLongBigList coefficients
      The displacement coefficients.

    • rank

      protected final SparseRank rank
      The sparse ranking structure containing the unused entries.

    • signatureMask

      protected final long signatureMask
      The mask to compare signatures, or zero for no signatures.

    • signatures

      protected final LongBigList signatures
      The signatures.

  • Constructor Details

    • CHDMinimalPerfectHashFunction

      protected CHDMinimalPerfectHashFunction(Iterable<? extends T> keys, TransformationStrategy<? super T> transform, int lambda, double loadFactor, int signatureWidth, File tempDir, ChunkedHashStore<T> chunkedHashStore) throws IOException
      Creates a new CHD minimal perfect hash function for the given keys.
      Parameters:
      keys - the keys to hash, or null.
      transform - a transformation strategy for the keys.
      lambda - the average bucket size.
      loadFactor - the load factor.
      signatureWidth - a signature width, or 0 for no signature.
      tempDir - a temporary directory for the store files, or null for the standard temporary directory.
      chunkedHashStore - a chunked hash store containing the keys, or null; the store can be unchecked, but in this case keys and transform must be non-null.
      Throws:
      IOException

  • Method Details