METHOD FOR PARSING INFORMATION OF DATABASES RECORDS USING WORD-LOCATION PAIRS AND METAWORD-LOCATION PAIRS

In a computer implemented method, a plurality of records are stored in a database at unique record addresses. The information of each record is parsed into a set of individual words where each word represents a portion of the information of a particular record. A unique sequential location is assigned to each word so that the location of a first word of a next record sequentially follows the location of a last word of a previous record. Pairs are formed from the words and their assigned location. For attributes common to a particular record, generate a record metaword, and assign the location of the last word of the record to the metaword. For attributes common to subsets of words, generate a first and second field metaword, and assign the location of the first and last word of the subset to the first and second field metaword to form pairs. The pairs are sorted in a memory, first according to the words and metawords, and second according to their locations so that each unique word and metaword is immediately followed by a sequential list of locations where the unique word and metaword occurs in the database.