International Business Machines Corporation

Julie Ann Watts

Kalpana Shyam

Robert William Lyle

James Zu-Chia Teng

San Yu Phoenix

Craig Alan Friske

Christopher John Crone

Karelle Lynne Cornwell

Margaret Ann Bernal

James Alan Ruddy

Donald J Haderle

Jay A Yothers

Paramesh Sampatrai Desai

Regina J Liu

James Z Teng

James Willis Pickel

Oliver Draese

Tanya Couch

Jerry Mukai

Binghui Zhong

Charles H Lin

Cherri M Vidmar

George F Wang

Shirley Zhou

Namik Hrle

Curt Lee Cotner

Fen-Ling Lin

Yao-Ching Stephen Chen

John Marland Garth

Akira Shibamiya

Steven John Watts

Christina Marie Lee

Claire Willey McFeely

Laura Michiko Kunioka-Weis

Yao Ching Stephen Chen

Andreas Maier

Charles Roy Bonner

Michael R Shadduck

Gary Howard Sockut

Yumi Kimura Tsuji

Eric N Katayama

Peter Hoa

Stephen Walter Turnbaugh

Stephen Yao Ching Chen

Georgia Carroll Fuller

Kathryn Ruth Zeidenstein

Robert J Hollines, Iii

William Y Kyu

Hsuiying Yen Cheng

RamanaKumari M Croisettier

Daniel Keith Courter

Jennifer Wha-Yuh Ou

James Zu-Chia Tseng

David Elmer Raiman

Kevin Daniel Gougherty

Richard Vincent Becraft, Jr

Robert William Engles, Deceased

Barbara Tanya Couch

Claire Louise McFeely

Jerry Wayne Goldsmith

In a high insert throughput in a database system, the database management system (DBMS) caches, asynchronously from data insert transactions, a list of storage location identifiers for storage locations in a table space in the database system. When a request for an insert transaction with an insert size is received from an agent, the DBMS identifies a storage location on the list with an amount of free space meeting the insert size. The DBMS removes the storage location identifier from the list and returns the storage location identifier for the storage location. When the insert transaction completes, the DBMS determines a remaining free space at the storage location and adds the storage location identifier to the list according to the remaining free space. By removing storage location identifiers from the list, another agent cannot receive the same storage location identifier for a parallel task. Thus, space contention is avoided.

Method and technique of achieving extraordinarily high insert throughput

Dropping of columns from a table with data availability, where the columns in the table are each associated with a column number, includes: executing a statement to drop a given column in the table and deferring an application of the statement until a reorganization of a current data set including the table. When the reorganization of the current data set is performed, the reorganization includes: updating column numbers for columns in each row of the table using a mapping data structure to remove the given column; loading the plurality of data rows with the updated column numbers into a shadow data set; applying to the shadow data set any changes to the table that are concurrent with the reorganization; updating a schema definition of the table with the updated column numbers; and switching the current data set to the shadow data set.

Dropping columns from a table with minimized unavailability

A system for controlling concurrency of access to data in a database system is provided. The system includes receiving a lock request for access to data in the database system, in which the lock request is a request for a page lock or a row lock for a corresponding row or page in the database system containing the data. The method further includes identifying a partition in the database system that contains the row or the page in the database system containing the data; associating the lock request with a partition lock on the partition that contains the row or the page in the database system containing the data; and accessing the data using the partition lock.

System for controlling concurrent access to data in a database system

A method for generating an execution plan for updating and retrieving data from a database in a single process includes: receiving a statement by a server to update a database with a first set of data and to retrieve a second set of data from the database; building a first execution plan to update the database with the first set of data; building a second execution plan to retrieve the second set of data from the database; and building a single execution plan including a combination of the first and second execution plans. The single execution plan allows for the updating and retrieval of data with a single crossing of an interface between a client and the database. By performing both functions in this manner, the efficiency of performing such tasks on the database is significantly increased.

Method of returning data during insert statement processing

A method for selectively releasing locks when rolling back to a savepoint includes: providing at least one savepoint in a transaction, where the at least one savepoint and the transaction are separate lock owners, where at least one lock is assigned to the at least one savepoint and at least one lock is assigned to the transaction; rolling back the transaction to the at least one savepoint; and releasing any locks owned by the at least one savepoint, where any locks owned by the transaction is maintained. Ownership of locks to persist until commit is assigned to the transaction. Ownership of locks to be released when rolled back to a savepoint is assigned to the savepoint. When roll back to a given savepoint occurs, locks owned by the savepoint are released while locks owned by the transaction are maintained.

Method for releasing update locks on rollback to savepoint

A method and system for controlling concurrency of access to data in a database system, includes: partitioning a table in the database system into a plurality of partitions; receiving a request for access to data; determining a partition of the plurality of partitions that contains the data; determining if the data has been committed; and if so, avoiding locking the partition in response to the request. By avoiding locking the partition when the data has been committed, the number of partition locks that need to be requested from a local resource lock manager is reduced, improving performance.

System and method for selective partition locking

Provided are a method, system, and program for making data available to an application program. A cursor is generated positioned with respect to a result table, wherein the cursor specifies a search criteria, and wherein the result table includes rows from a base table that satisfy the search criteria. A fetch request is received indicating to position the cursor on a plurality of rows of the result table. The cursor is positioned on the plurality of rows of the result table indicated in the fetch request that satisfy the search criteria.

Method, system, and program for positioning a cursor on rows of a result table

The present invention is related to a method and system for preserving an original table schema for a table in a database system that supports dynamic table schema changes. The method and system includes storing the original table schema for the table in a designated table prior to performing a schema change on the table. By storing the original table schema in the designated table, a backup copy of the table that includes rows which are not self-describing can be used for data recovery.

Method and system for preserving an original table schema

Provided are a method, system, and program for managing a database. A request is processed to create an index on at least one key column of a table in a computer readable medium. A determination is made of table partitions in which rows of the table are stored. One index partition is generated for each determined table partition, wherein each index partition includes nodes, wherein each node in each index partition includes a key value from at least one column in a corresponding table row in the table partition associated with the index partition and a location identifier identifying the corresponding table row in the corresponding table partition. In certain embodiments, a technique is implemented during queries to limit the search to a subset of the multiple index partitions when possible and return rows in the order of the key columns of the index while the same range of key values can exist across multiple index partitions.

Method, system, and program for managing database operations with respect to a database table

A method, apparatus and article of manufacture, implementing the method, for operating a database management system stores information in at least one entry. Each entry is associated with a row. At least one entry comprises one or more columns. At least one of the columns has a varying length. At least one entry comprises an offset array having one or more offsets associated with the columns having the varying length. In another aspect of the invention, a column extraction table is built and used, with the offset array, to access a specified column of an entry. Alternately, code is generated to access a specified column based on the offset array of an entry. In yet another aspect of the invention, a comparison table is built and used, with the offset array, to compare subsets of the entries.

Database management system using offsets in entries with at least one varying-length column

According to the method, apparatus, and computer readable medium of the present invention, a scan is performed by accessing all rows of a table in a relational database, evaluating each row to determine whether the row satifies a set of predicates of a database query, and returning the row if it satisfies the set of predicates. According to a preferred embodiment of the present invention, all rows are evaluated regardless of current locks. The scan is continued if the row does not satisfy the set of predicates regardless of the current lock on the row. The locking semantics according to the invention has the advantage that it completely removes the problem of lock contention on the rows that do not satisfy statement predicates.

Method and system for reduced lock contention in SQL transactions

Traditional update semantics are provided when updates change the location of records in a database. The system detects that an update to a record will cause the record to change locations in a database from an original location to an updated location. The system selectively prevents activity in the database between the original location and update location until the updated record has changed locations in the database from the original location to the updated location.

Providing traditional update semantics when updates change the location of data records

System, method, and program product for managing transactions on a plurality of Database Management Systems residing on separate participant servers and a coordinator server. A single transaction proceeds by multiple steps across more then one of the plurality of Database Management Systems as participants. Savepoints are assigned to the participants across the Database Management System, with at least one savepoint within a multiple step transaction. The DBMS responds to application requests to rollback to savepoint by rolling back the transaction, across the Database Management System participants, to the savepoint Further disclosed is a program product that contains code, which, either as written on the medium, or as instantiated in RAM, or both, controls the operation of a distributed database management system.

Method for managing distributed savepoints across multiple DBMS's within a distributed transaction

Growing community of inventors

Morgan Hill, CA, United States of America

Julie Ann Watts