When searching for a pattern in a string, we refer to the pattern as the needle and the whole corpus as the haystack. The Horspool string search algorithm implemented in this recipe performs well for almost all pattern lengths and alphabet sizes, but is ideal for large alphabet sizes and large needle sizes. Empirical benchmarks can be found by navigating to the following URL:

http://orion.lcg.ufrj.br/Dr.Dobbs/books/book5/chap10.htm

By preprocessing the query, the algorithm is able to efficiently skip redundant comparisons. In this recipe, we will implement a simplified version called Horspool's algorithm, which achieves the same average best case as the Boyer-Moore algorithm, benefits from having a smaller overhead cost, but may in very rare circumstances suffer the same worst-case running time as the naive search when the algorithm performs too many matches. The Boyer-Moore algorithms should only be used if the extra prepossessing...