Full-text search apparatus utilizing two-stage index file to achieve high speed and reliability of searching a text which is a continuous sequence of characters

A text search apparatus operable for deriving a set of text position values expressing all positions of occurrence of a search string within an object text, said search string being an arbitrarily specified string of characters from a predetermined character set, said object text being formed of a finite number of characters from said character set, the apparatus comprising:

• a dictionary list having registered therein a predetermined list of dictionary words or identifiers of said dictionary words, each of said dictionary words being a string of characters from said character set,
• text position table means comprising a plurality of text position elements each expressing a combination of one of said dictionary words which appears within said object text, or an identifier of said dictionary word, in conjunction with a text position value of said dictionary word, each dictionary word of a text position element in said text position table being an extension word which is a maximum length word within a set of dictionary words each extending from a leading character position of said text position element;
• extension word position correspondence table means having registered therein information expressing, for each of respective partial character strings of said extension words, a corresponding set of extension words and respective positions of said partial character string within each of said corresponding extension words,
• extension word assessment means for obtaining from said extension word correspondence table means a first set of extension words each containing said search string as a partial character string and intra-word position information indicating respective positions of said search string within the words of said first set of extension words,
• text position element look-up means for obtaining first sets of text position elements from said text position table means, each set of said first sets of text position elements being keyed to a corresponding extension word of said first set of extension words,
• total set formation means for operating on said first sets of text position elements in conjunction with said intra-word position information, to obtain a first set of text position values of said search string, and
• cover word assessment means for determining a specific sequence of cover words extending from a leading character to a final character of said search string, each of said cover words other than a final cover word of said sequence being a dictionary word, said final cover word being a dictionary word or a leading partial character string of a dictionary word,
• wherein said extension word assessment means obtains from said extension word position correspondence table means respective second sets of extension words, said sets corresponding to respective ones of said cover words, and obtains intra-word information expressing respective positions of each of said cover words within each of the extension words of the corresponding one of said second sets,
• said text position element look-up means obtains from said text position table means, for each extension word of each set of said second sets of extension words, a corresponding second set of said text position elements,
• said cover word assessment means executes comparison processing of said second sets of text position elements to select groups of sequentially occurring text position elements, each of said groups containing said cover word sequence extending continuously from within a leading text position element of the group,
• and wherein said total set formation means obtains, based on respective leading text position elements of said groups in conjunction with said intra-word information from said extension word assessment means, a second set of text position values of said search string, and combines said second set with said first set of text position values to obtain a complete set of text position values of said search string.