Did you know ... | Search Documentation: |
Pack hashtbl -- README.md |
This library provides two implementations of hash tables in Prolog. Hash tables store key-value pairs and provide predicates to add, look up, remove entries as well as conversions to and from lists and higher order predicates for mapping, folding, and iterating over hash tables. The hash table variants differ in their approach to purity, i.e., what kind of side effects their implementation uses.
The nb_hashtbl
module provides destructive ("non-backtrackable") hash
tables:
:- use_module(library(nb_hashtbl)). ?- empty_nb_hashtbl(T), nb_hashtbl_set(T, x, 1), nb_hashtbl_set(T, y, 2), nb_hashtbl_get(T, x, X). T = ..., X = 1.
You can add the same key multiple times with put
, keeping previous values,
or use set
to replace the last stored value:
?- empty_nb_hashtbl(T), nb_hashtbl_put(T, x, 1), nb_hashtbl_put(T, x, 2), nb_hashtbl_to_list(T, Pairs). T = ..., Pairs = [x-2, x-1]. ?- empty_nb_hashtbl(T), nb_hashtbl_put(T, x, 1), nb_hashtbl_set(T, x, 2), nb_hashtbl_to_list(T, Pairs). T = ..., Pairs = [x-2].
The p_hashtbl
module provides non-destructive ("pure") hash tables:
?- list_to_p_hashtbl([x-1, y-2], T0), p_hashtbl_get(T0, x, X). T0 = ..., X = 1. ?- list_to_p_hashtbl([x-1, y-2], T0), p_hashtbl_delete(T0, x, T1), p_hashtbl_get(T1, x, X). false.
Both variants provide higher-order map
, fold
, and iter
operations:
hashtbl_sum(_Key, Value, Accumulator, Result) :- Result is Value + Accumulator. ?- list_to_p_hashtbl([x-3, y-2, z-1], T), p_hashtbl_fold(T, hashtbl_sum, 0, Sum). T = ..., Sum = 6.
All of the hash tables defined in this library use term_hash/2 to hash a key that can be associated with one or more values in the table. Due to the requirements of term_hash/2, the key must always be ground. There are no restrictions on the stored value.
Depending on which operation is used, adding a key-value pair to a table that already contains that key may shadow a previous entry. Subsequent get or delete operations will refer to the newest entry; if it is deleted, earlier shadowed entries become visible again.
The implementations avoid copying the keys and values. Later variable bindings or destructive updates to terms may thus affect the stored values.
The internal representation of the hash tables is subject to change as features are added. Writing hash tables to files and reading them back might not work between different versions of this library, or even with the same version once randomization is implemented. If you want to serialize a hash table for I/O, you should convert it to a list first and then convert the list back to a hash table.
The non-backtrackable hash tables are resized automatically when their load exceeds a certain threshold.
The implementation of the pure hash tables is quite naive, and a bad choice of the number of buckets leads to terrible performance. However, dynamic resizing as for the non-backtrackable hash tables performs even worse because hash tables with more buckets become more expensive to copy every time an element is inserted. Some sort of hierarchical system is needed to solve this.
All the predicates should check that their hash table arguments are indeed instantiated to hash tables and throw appropriate instantiation errors. Currently they may fail silently or instantiate uninstantiated arguments.
We might consider allowing other hashes than term_hash/2. In particular, hashing with randomization should be useful. Randomized hash tables are intended to protect against denial-of-service attacks by attackers who can engineer a large number of collisions. The standard libraries of Python and OCaml use randomized hashing, for example.
This library requires SWI-Prolog; it was tested with SWI-Prolog 6.6.4 and should work with newer versions.
pack_install(hashtbl).
The pure part should be reasonably portable to other ISO Prolog systems.
This library was written by Gergö Barany <gergo@tud.at>. It lives at https://github.com/gergo-/hashtbl.
Comments, questions, bug reports, feature requests, and patches are welcome!
This library is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public License along with this library (file LICENSE; see also LICENSE.GPL for the GNU General Public License). If not, see http://www.gnu.org/licenses/.