Byte codes are a practical alternative to the traditional bit-oriented compression approaches when large alphabets are being used, and trade away a small amount of compression effectiveness for a relatively large gain in decoding efficiency. Byte codes also have the advantage of being searchable using standard string matching techniques. Here we describe methods for searching in byte-coded compressed text and investigate the impact of large alphabets on traditional string matching techniques. We also describe techniques for phrase-based searching in a restricted type of byte code, and present experimental results that compare our adapted methods with previous approaches.
展开▼