Abstract: A stream data processing system capable of highly reliable and highly usable real time data processing, is provided with: (1) a mechanism of copying some or all of stream data, and archiving the stream data in a non-volatile memory to allow real time data and archive data to be used seamlessly; and (2) a mechanism of improving the performance of query processings by using a plurality of stream data processing systems cooperatively.

Abstract: In order to realize a practical data warehouse system with the use of a network in a distributed environment, for example, data collectors for creating/managing each replica of server data for clients shown in FIG. 1 are connected to a network. A data collector holds a data request condition given from a user via a client and negotiates with a server which supplies data according to such resources as a storage unit, a CPU, etc., usable by the data collector, thereby the range and quality of the data for which a replica is to be created can be adjusted. For query processing, a part or the whole of a query from a client is processed with the use of a replica created in the data collector. Queries which cannot be processed with any replica in the data collector are transferred to the server. Because a replica created in the data collector can be used to process a query from a client in such a way, the server load can be reduced.

Abstract: In order to realize a practical data warehouse system with the use of a network in a distributed environment, for example, data collectors (101 and 102) for creating/managing each replica of server data for clients (103 and 104) as shown in FIG. 1 are connected to a network (113). A data collector holds a data request condition given from a user via a client and negotiates with a server (114, 122) which supplies data according to such resources as a storage unit, a CPU, etc. usable by the data collector, thereby the range and quality of the data for which a replica is to be created can be adjusted. For query processing, a part or the whole of a query from a client is processed with the use of a replica (123) created in the data collector. Queries which cannot be processed with any replica in the data collector are transferred to the server. Because a replica created in the data collector can be used to process a query from a client device such a way, the server load can be reduced.

Abstract: By corresponding the external databases to the predicates used in query processing according to a predetermined ontology, it is allowed to issue a query through the ontology. The query thrown into the integrated database is expanded into the set of approximate queries according to approximation rules provided in the ontology. The optimization module refers to the query capability of the external databases to perform the query optimizing in which these approximate queries can be executed efficiently as a whole.

Abstract: A method for integrating data, stored not only in a relational database management system but also in data sources such as legacy application programs or tertiary storage where inquiry cannot be executed efficiently, and for performing inquiry quickly. A distributed index for use by a data source is generated in a database hub in advance, and a distribution index application unit converts and disassembles an inquiry using the distributed index.

Abstract: By introducing the concept of negation into an association rule, the number of combinations of items, serving as candidates of the association rule containing negative items, becomes enormous for mining. The invention sorts out frequent itemsets to be included in an association rule in advance by ensuing its upper bound of confidence satisfying conditions of a confidence and a confidence increment input by an user. Also, by checking a logic product for each bit of a bit string reflecting the presence of each item in the itemset, the combinations of items as candidates are pruned. The pruning of unnecessary combinations is effectively carried out. The number of combinations of the items for compiling association rules is reduced.

Abstract: The present invention improves the scalability of application servers and database servers in Web systems, thereby providing higher cost-performance Web systems. The session state of an application is stored in a clustered cache database as required. This enables another application running on an arbitrary cluster server to read the stored session state from the clustered cache database to continue the processing of the application. In addition, a cache database cluster in which data can be updated on the cluster servers is configured, thereby providing a mechanism of causing them to synchronize with the database servers as necessary. Furthermore, a mechanism of adding cache databases to the cache database cluster, thereby enabling the system to respond to an increase in database access load.

Abstract: Method and apparatus for query processing including an optimization controller, responding to a query from an application related to an database, and controlling optimization processing of the query. A query classification definition and a query operation direction associated with the query classification definition are defined on a storing means. The query classification definition classifies the queries according to classifications and attributes including environment for use, time for use, database to be accessed, and the amount of data that is processed. The query operation direction defines strength of the direction including “Force”, “Recommended”, “Neutral”, and “Deny”, which is used to select operation at the time of query execution from several alternatives.

Abstract: For efficiently executing a join in a star schema, a virtual concatenate indexes are stored in a database for defining combinations of a plurality of indexes including at least one of indexes for retrieving corresponding records from a column value on a fact table and one of indexes for retrieving corresponding records from a column value on a dimension table. Indexes indicated by the corresponding virtual concatenate index are sequentially accessed for processing a query which involves a join of the tables. Before processing the query, the virtual concatenate index is materialized. Specifically, a plurality of indexes specified by the virtual concatenate index are joined only within a specified range of column values to point associated records on the fact table, and a set of record IDs of the pointed records are stored corresponding to the column values.

Abstract: A query processing system in which a virtual table to enable multiple mapping to a plurality of databases is provided in a data processing system to receive a query from an application program and partial replicas as copies of part of the data in the virtual table is prepared in storage so that the whole or part of a query can be processed using the partial replicas. For query units which cannot be processed using the partial replica, means to select an appropriate distributed join method and optimization to push down query processing to real databases as far as possible are used. The system allows transparent access to multiple different real databases according to access conditions without changing the column to be accessed in the virtual table or switching the virtual table, thereby eliminating the need to create a new application for each newly developed service.

Abstract: In a file transfer system, a file stored in a server computer is transmitted to a client computer through a network. A portable mobile apparatus also connectable to the network has a transmission requesting means for requesting the server computer that the user of the mobile apparatus be authenticated and a desired file be transmitted from the server computer to the client computer. The server computer has a communication control unit for transmitting the desired file requested by the mobile apparatus to the client computer after authenticating the user of the mobile apparatus.

Abstract: A data request analysis unit for accepting and analyzing a request from a client computer, a data conversion program generation unit, and a data conversion unit are provided in a magnetic disk apparatus, the data conversion program generation unit downloads a program module by way of a network as required based on the analysis result obtained by use of the data request analysis unit to generate a data conversion program, and the data conversion unit downloads the conversion program, and converts the type of the data read out from the magnetic memory medium by use of the program, and directly transfers the converted data to the client computer by way of a network interface, thereby, the present invention provides a general purpose magnetic disk apparatus which is capable of flexibly accommodating the request from the client computer, the scalability of the whole system is improved because the data is converted in the magnetic disk apparatus, and the process efficiency of the client computer is improved.

Abstract: By introducing the concept of negation into an association rule, the number of combinations of items, serving as candidates of the association rule containing negative items, becomes enormous for mining. The invention sorts out frequent itemsets to be included in an association rule in advance by ensuing its upper bound of confidence satisfying conditions of a confidence and a confidence increment input by an user. Also, by checking a logic product for each bit of a bit string reflecting the presence of each item in the itemset, the combinations of items as candidates are pruned. The pruning of unnecessary combinations is effectively carried out. The number of combinations of the items for compiling association rules is reduced.

Abstract: The present invention improves the scalability of application servers and database servers in Web systems, thereby providing higher cost-performance Web systems.

Abstract: By corresponding the external databases to the predicates used in query processing according to a predetermined ontology, it is allowed to issue a query through the ontology. The query thrown into the integrated database is expanded into the set of approximate queries according to approximation rules provided in the ontology. The optimization module refers to the query capability of the external databases to perform the query optimizing in which these approximate queries can be executed efficiently as a whole.

Abstract: A data relay server connected between a client computer and a database server produces a query message obtained by encrypting a retrieval condition designated by the client computer, and transmits the query message to the database server. The database server encrypts a specific data item which is designated by the retrieval condition and read out from the database, retrieves service information matched with the retrieval condition by encrypted data matching, and replies the service information as a retrieval result to the data relay server.

Abstract: In order to realize high speed process for grouping the records having the identical values of one or more columns of the input list, the input list is canned, a hash value is generated using a hash function in which a value of the column as the non-vacant partial aggregation of the columns for the grouping is used as the argument, and it is determined whether two or more records having the hash values exist or not. The input list is scanned again and the calculating process of the aggregation columns is immediately executed for the records having the hash value for which it is determined by the first scanning that there is only one record having the identical hash value, the result of such calculating process is output and the records are defined as the input of the ordinary grouping process for the records which are determined to exist as the two or more records.

Abstract: A method for integrating data, stored not only in a relational database management system but also in data sources such as legacy application programs or tertiary storage where inquiry cannot be executed efficiently, and for performing inquiry quickly. A distributed index for use by a data source is generated in a database hub in advance, and a distribution index application unit converts and disassembles an inquiry using the distributed index.

Abstract: A query issue processing method, a query conversion processing method, and a data control processing method are provided for enhancing the efficiency of random sampling processing for use in a database processing system. In query issue processing 2, a query including random sampling processing is issued. In query conversion processing 8, application sequences of random sampling processing and another query processing are exchanged by considering a sampling unit of the random sampling processing. Further, in record control processing 4, random access to a secondary storage device is reduced, thereby enhancing random sampling processing efficiency. Unlike the conventional query conversion processing not considering the sampling unit, the issuance of the query including random sampling processing and performing query conversion by considering the sampling unit allow random sampling to be applied also to a query including aggregation processing, thereby enhancing the efficiency of queries in a wider range.