Hashing with Linear Probing and Frequency Ordering

Linear probing of a scatte r (or hash) table interpre ts each key or item (these terms are inte rchangea ble here) as a probe index into the table [1]. 1 Typically, a key is divided by the table size and the remainder is used for indexing. If the selected slot is empty, the item is not present. Should the slot contain some other key, eacl1 next highe r location is checked until the item is found , an empty slot is discovered , or the whole table has been examined. (Indexes that exceed table sizes wrap around.) Table slots thus sea rched defin e a key's collis ion resolution sequence. Linear probing is generally not suitable for nearly filled tables, since as empty slots disappear, sea rches get very long. Nonetheless, linear probing can be improved by allowing for item frequ ency-of-reference.