Abstract: A method and apparatus enables partitioning of data to efficiently store and retrieve data relating to a database system (either a parallel or non-parallel database system). Such partitioning may be performed by receiving information associated with at least one characteristic of the data and performing an algorithm to divide the data into related data segments. The data segments may be stored in a distributive database system based on the characteristic associated with the data.

Abstract: A database system for selecting rows from a partitioned database table is disclosed. The partitioned database table includes rows and columns and is divided into partitions with at least one of the partitions in the table being populated by one or more rows. The system includes one or more nodes, each of the one or more nodes providing access to one or more CPUs. Each of the one or more CPUs provides access to one or more virtual processes. Each process is configured to manage data, including the partitioned database table, stored in one of a plurality of data-storage facilities. The system also includes a partitioned table access component configured to select rows from the table by creating a file context, which stores at least location data for a row and a first value associated with the row, for each populated partition, determining the lowest first value stored by the file contexts, and identifying rows with a particular first value by reading the file contexts.

Abstract: A stratified sampling mechanism is provided in a database system. The stratified sampling mechanism includes defining a clause in a query that indicates stratified sampling is desired. Data from a source table is stratified into different subgroups based on stratification conditions in the query. Sampling is performed within each subgroup.

Abstract: A computer-implemented data mining system that analyzes customer transaction data using Factor Analysis/Retail Data Mining Segmentation. The data is accessed from a relational database, and then a factor analysis function is performed on the data to create a factor loadings matrix that has factors as columns and observed variables from the customer transaction data as rows, wherein each of the observed variables is assigned to one of the factors in the factor loadings matrix that has the maximum value for the row. New variables are derived by means of a factor-scoring method that combines the variables into the factors in the factor loadings table. Customer destination segments are identified from the relational database using the factors. Additional customer destination segments are identified by means of a clustering tool using the derived new variables.

Abstract: A database system includes a storage that contains plural tables as well as a predefined data structure. The database system is able to, in response to a join query, perform a join of two or more tables. The database system also is able to determine, based on values contained in the predefined data structure, whether a group-by operation can be skipped.

Abstract: A method and apparatus is provided in a parallel database system having a plurality of nodes for computing a moving aggregate of an attribute of a relation having multiple tuples. Portions of the relation are distributed across the plurality of nodes of the database system. For each node i, the database system identifies one or more other nodes that contain tuples of the relation which are covered by a moving window of each tuple of the relation at node i. For each such identified node, a value representing an aggregate of the tuples at each such identified node is communicated to node i. The moving aggregate is then calculated using at least tuples of the relation at node i as well as the value from each of the identified one or more nodes.

Abstract: In a database system, when an operation (a request or a transaction) that modifies one or more rows of a table is performed, multiple images of the row that is being modified (e.g., added, deleted, or updated) is stored in the table. State identifiers are provided to identify the state of each row, such as a “before image” row and an “after image” row. A before image row represents a row before a modification or “mutation” operation is performed. An after image row represents the latest row as processed by a current modification operation. A before image row is retained in each table as long as the transaction that created or updated the row is active. As soon as the transaction commits successfully, the before image rows become reusable and can be reused by another request or transaction. The state identifier associated with each row can also indicate that a “mutation” is not being performed on a row; that is, an operation that modifies the contents of the row is not currently being performed.

Abstract: Methods, systems, and executable instructions are provided for selecting an optimal tree summary operation. Tree summary operations are evaluated based on their efficiencies in calculating a tree summary from a tree having a plurality of nodes. Moreover, in one embodiment, by using tree constants and processing constants the efficiencies of the tree summary operations are determined. Once the efficiencies of the tree summary operations are determined, one of the tree summary operations is selected as an optimal tree summary operation for calculating the tree summary of the tree.

Abstract: A circuit board includes multiple signal layers, in which signal lines are routed, and power reference plane layers, in which power reference planes (e.g., power supply voltage or ground) are provided. Vias are passed through at least one signal layer and at least one power reference plane layer, or alternatively, vias are passed through at least two power reference plane layers. In one arrangement, a first clearance is defined around the via at the signal layer and a second clearance is defined around the via at the power reference plane layer. The second clearance is larger in size than the first clearance to match or tailor the impedance of the via as closely as possible with the impedance of the signal line that the via is electrically connected to.

Abstract: A customer relationship management system for storing and managing information for an E-Business retailer. The customer relationship system includes a database system for storing and organizing said information, a logical data model defining the manner in which said information is stored and related within the database system, and a subject area within the logical data model defining the manner in which profile information concerning web site visitors to a web site operated by the E-Business retailer is stored and organized within the database system. Visitors may include any individuals, households or organizations that are of interest to the E-Business retailer.

Abstract: Methods and apparatus for streaming data from a database system to a client system. In one implementation, a database management system includes: one or more data storage facilities for use in storing data composing records in tables of a database, where the data includes media data stored in one or more media objects; one or more processing modules configured to manage the data stored in the data-storage facilities; one or more quality of service modules configured to manage streaming media data from the data-storage facilities; and a database management component configured to receive stream requests and database queries for data stored in the data-storage facilities, including passing a stream request indicating media data to the processing module that corresponds to the data-storage facility storing the media data indicated by the stream request.

Abstract: A method for reorganizing rows from a partitioned database table is disclosed. The partitioned database table includes a plurality of populated partitions. Rows are organized in each of the populated partitions in accordance with a first value associated with each row. A file context is created for each partition of a subset of the populated partitions, each file context stores at least location data for a row in the partition and the first value associated with the row. Rows and merged from the subset of partitions into a single first-merge partition in order of the first value associated with each row. The creation of file contexts and merging of rows in the corresponding partitions is repeated until the subsets have included all populated partitions.

Abstract: A parallel database system includes multiple processing units and associated storage modules. To achieve even distribution of workload, rows of a table are distributed as evenly as possible across the multiple processing units using a predefined partitioning algorithm. The predefined partitioning algorithm defines groups and probabilities assigned to boundary points defining the groups. Partitioning of data is based on the boundary points and the assigned probabilities.

Abstract: A method, computer program and system for joining two tables, t1 and t2 are described. Each table contains rows and columns and is divided into one or more partitions. A correlation function is calculated between a first correlated value column of table t1 and a second correlated value column of table t2. Following a query request, a joining algorithm is used to join t1 and t2, using a reduced set of partitions based in part upon the correlation function.

Abstract: A method for reading a multiple byte data element stored in both first and second memories. Selected bytes of the data element are invalidated in the first memory. Valid bytes from the first memory are combined with remaining bytes from the second memory in response to a read request.