IndexHash

A simple HashMap for C.

Building

To build the basic test platform, run:

$ make

The library can be used with map.h and map.o, if exported. Simply export to a lib directory in your project, point your makefile to the .o file, and include the map.h file in the include directory.

Using

To use the map, initialize it with:

Map* map = newMap();

This creates a new Map instance.

All actions of the map can be done using the provided functions in the Map's struct. All functions require that the user pass in the map instance as the first argument.

map->add(map, key, value);
map_>remove(map, key, dataDelete);
map->get(map, key);
map->set(map, key, newValue);
map->delete(map, dataDelete);

Values are void*, and keys are charact4er strings. Keys do not need to be allocated by the user, and a character array will suffice, as long as it is a valid string. The system will allocate its own internal keys.

When setting data to new values, it is recommended to retrieve the data beforehand, as it will not bve freed by the system, that is left to the programmer in this instance.

The dataDelete flag is a notice for the map, when deleting an item or itself, to call free() on each data filed in the map. If you have malloc'd items in the map, and cannot free them yourself, set this flag to 1. Otherwise, keep it at 0.

Data Representation

In a move that I may or may not regret, data is stored in the map as a void*. This is done to keep the system as flexible as possible, and to let it handle any sort of object. However, this also means that it cannot store un-malloc'd data points, unless you do something squirrelly. (You can store literals, but there will be no type checking, and are a maximum of 64 bits in size). This is technically undefined behavior (though I myself have done it successfully).

Organization

Build as a "box of bins", the map uses a series of bins, each which is a linked list. When a given bin reaches capacity, then the system will increase the number of bins, and redistribute the nodes. This is an expensive process, but should not happen too often.

Hashing

It uses a hash algorithm used by sdbm via http://www.cse.yorku.ca/~oz/hash.html.