Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
1998-08-03
2001-11-06
Black, Thomas (Department: 2171)
Data processing: database and file management or data structures
Database design
Data structure types
C707S793000
Reexamination Certificate
active
06314418
ABSTRACT:
BACKGROUND OF THE INVENTION
(1) Field of the Invention
The present invention relates to an index managing (controlling) unit, an index updating method, an index managing method, a computer-readable recording medium retaining an index updating program, and a computer-readable recording medium holding an index managing program, which are suitably employed for managing and updating an index file for use in a retrieval system which retrieves a large volume of data information using an item as a key (on the basis of an item) and extracts related information.
An index file is for the purpose of storing a large amount of record information in relation to key information, and particularly, an index file being of an inverted file type is for searching record information at a high speed on the basis of key information forming an item which organizes an index, and is available for full-text retrieval.
(2) Description of the Related Art
A common retrieval system is allowed to conduct an information retrieval with respect to, for example, a group of document files. In more detail, the retrieval system is designed to perform such information retrieval processing as to use some word as a key to output a group of document numbers, associated with this key, as a retrieval result concurrently with retrieving a group of document files with document numbers.
In such retrieval processing, in order to speed up the processing to be taken until the output of a retrieval result after the input of a retrieval key, the management of information about a group of document files is made in a manner that held on a storage area is an inverted file type index file being the collection of retrieval results corresponding to keys determined in advance.
Briefly, when one key is inputted to the retrieval system, a retrieval result can be outputted through only an operation of merely opening the corresponding record information in the above-mentioned index file on the basis of this key.
Meanwhile, for the area allocation on a storage area of an index file in a prior retrieval system, an area block with a given size is initially given as an area for a portion of record information corresponding to key information and the record information is stored therein, whereas record information exceeding the given size is recorded across a plurality of area blocks.
More specifically, an area block with a given length is allocated as an initially set value onto a storage area in connection with each key so that the record information corresponding to each key is put in that area block, and if the area size of the allocated area block is insufficient to the record information to be stored, an area block for storing the record corresponding to that key is given at a location remote from the initially allocated area and a chain is established between these area blocks, thereby securing the area for storing the record information corresponding to the key information.
However, in the case of the prior area allocation technique on the storage area in the index file, when applying as the index file the inverted file type for use in a full-text retrieval, the record information to be stored in conjunction with the key information significantly differ in size from each other depending upon the key information. In such a condition, if an area block with a given size is initially allocated in units of K bytes, a problem will arises in that, because the most keys only require a considerably smaller area than this area block, an excessive area consequently takes place for the storage of the record information, which can interfere with the effective utilization of the storage area.
On the other hand, in the prior retrieval system, if a constant area is allocated even in incrementing the areas, the increment expectable area sizes vary in accordance with the keys. Particularly, a serious problem here is that an excessive area is allocated with respect to a small-increment expectable key.
SUMMARY OF THE INVENTION
The present invention has been developed in order to eliminate the above-mentioned problems, and it is therefore an object of this invention to provide an index managing unit, an index updating method, an index managing method, a computer readable recording medium retaining an index updating program, and a computer readable recording medium retaining an index managing program, which are capable of effectively allocating areas for an index having a long record string for a key as seen in a full-text retrieval system or the like and further having a data structure, where the lengths of the record sections depend upon keys to extremely differ from each other, through the use of statistics information on an index, such as an occurrence frequency of record information for each key and others.
For this purpose, in accordance with the present invention, an index managing unit, which is made to control a storage condition of an index composed of key information forming items and content information forming the contents of the key information on a storage area, comprises an index storage managing section for managing, on the storage area, the content information corresponding to the key information constituting the index and a management information storage managing section for managing, on the storage area, management information associated with the index, wherein the management information storage managing section includes a management information updating section for updating the management information undergoing the management on the storage area on the basis of record information constructed in units of a set of key information and content information inputted as information to be added to the index managed on the storage area, while the index storage managing section includes an area calculating section for calculating a free area, to be set when necessary, through an arithmetic or operation on the basis of the management information under the management by the management information storage managing section when the index on the storage area is updated on the basis of the record information inputted and an area allocating section for allocating a storage area for storage of the index on the basis of the free area calculated by the area calculating section.
Thus, in the index managing unit according to this invention, the area calculating section calculates a free area, to be set when needed, through an arithmetic on the basis of management information under management by the management information storage managing section while the area allocating section allocates a storage area for storage of the index on the basis of the calculated value, which can allocates an area with an appropriate size in accordance with the management information of the record information inputted while preventing an excessive area allocation.
Furthermore, in accordance with this invention, an index updating method of registering information to be added to an index on a storage area to update the index composed of key information forming items and content information forming the contents of the key information which are managed on the storage area comprises: a management information updating step of receiving input of record information composed of a set of key information and content information corresponding to the key information as the information to be added to the index to update management information for the index on the basis of the inputted record information; an extraction step of extracting, on the basis of the key information corresponding to the record information, a position on the storage area accepting registration of the content information of the inputted record information; a decision step of deciding whether or not successive free areas assuming a size needed for the registration of the content information of the record information exist at the position on the storage area obtained in the extraction step; an area calculating step of, when a result in the extraction step shows no extraction of the position on the storage area accepting the
Black Thomas
Fujitsu Limited
Staas & Halsey , LLP
Wang Mary
LandOfFree
Index managing unit, index updating method, index managing... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Index managing unit, index updating method, index managing..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Index managing unit, index updating method, index managing... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2604766