Abstract: A system and method is provided for query processing comprises: creating an index of a database and ordering a set of index candidates from the index into a list based on a set of heuristic rules. A query defining a query path is then reduced into a list of single path expressions. Each index candidate is matched against the list of single path expressions according to the ordering of the index candidates. The matched candidate nodes are also verified to insure that they satisfy the query path.

Abstract: System and method embodiments are provided for adaptive vector size selection for vectorized query execution. The adaptive vector size selection is implemented in two stages. In a query planning stage, a suitable vector size is estimated for a query by a query planner. The planning stage includes analyzing a query plan tree, segmenting the tree into different segments, and assigning to the query execution plan an initial vector size to each segment. In a subsequent query execution stage, an execution engine monitors hardware performance indicators, and adjusts the vector size according to the monitored hardware performance indicators. Adjusting the vector size includes trying different vector sizes and observing related processor counters to increase or decrease the vector size, wherein the vector size is increased to improve hardware performance according to the processor counters, and wherein the vector size is decreased when the processor counters indicate a decrease in hardware performance.

Abstract: Techniques are provided for encoded data processing which allows for continuous data processing as encoded data changes. Data is decomposed into one or more blocks with each block containing at least one data record. At least one data record within a given block is encoded with a first encoding process selected from one or more encoding processes. The first encoding process is associated with the given data block. Techniques evaluate whether or not to implement an encoding change for a given block when updating a given data record in the given block. Responsive to the evaluation, the given block is re-encoded with a second encoding process. Responsive to the re-encoding, the association of the given block is updated. A map is formed to convert the given data record encoded with the first encoding process to the second encoding process so as to preserve comparative relationships of the given data record.

Abstract: According to one embodiment of the present invention, at least a portion of data from a first processing system is archived onto a second processing system based on partitions of the data. A query received at the first processing system is processed at the second processing system to retrieve archived data satisfying the received query in response to determining at the first processing system that the received query encompasses archived data. Embodiments of the present invention further include methods, systems, and computer program products for archiving and accessing data in substantially the same manner described above.

Abstract: Techniques for encoded data processing which allow for continuous data processing as encoded data changes. Data is decomposed into one or more blocks with each block containing at least one data record. At least one data record within a given block is encoded with a first encoding process selected from one or more encoding processes. The first encoding process is associated with the given data block. Techniques evaluate whether or not to implement an encoding change for a given block when updating a given data record in the given block. Responsive to the evaluation, the given block is re-encoded with a second encoding process. Responsive to the re-encoding, the association of the given block is updated. A map is formed to convert the given data record encoded with the first encoding process to the second encoding process so as to preserve comparative relationships of the given data record.

Abstract: At least a portion of data from a first processing system is archived onto a second processing system based on partitions of the data. A query received at the first processing system is processed at the second processing system to retrieve archived data satisfying the received query in response to determining at the first processing system that the received query encompasses archived data. Embodiments of the present invention further include methods, systems, and computer program products for archiving and accessing data in substantially the same manner described above.

Abstract: According to one embodiment of the present invention, at least a portion of data from a first processing system is archived onto a second processing system based on partitions of the data. A query received at the first processing system is processed at the second processing system to retrieve archived data satisfying the received query in response to determining at the first processing system that the received query encompasses archived data. Embodiments of the present invention further include methods, systems, and computer program products for archiving and accessing data in substantially the same manner described above.

Abstract: Techniques for encoded data processing which allow for continuous data processing as encoded data changes. Data is decomposed into one or more blocks with each block containing at least one data record. At least one data record within a given block is encoded with a first encoding process selected from one or more encoding processes. The first encoding process is associated with the given data block. Techniques evaluate whether or not to implement an encoding change for a given block when updating a given data record in the given block. Responsive to the evaluation, the given block is re-encoded with a second encoding process. Responsive to the re-encoding, the association of the given block is updated. A map is formed to convert the given data record encoded with the first encoding process to the second encoding process so as to preserve comparative relationships of the given data record.

Abstract: Techniques are provided for encoded data processing which allows for continuous data processing as encoded data changes. Data is decomposed into one or more blocks with each block containing at least one data record. At least one data record within a given block is encoded with a first encoding process selected from one or more encoding processes. The first encoding process is associated with the given data block. Techniques evaluate whether or not to implement an encoding change for a given block when updating a given data record in the given block. Responsive to the evaluation, the given block is re-encoded with a second encoding process. Responsive to the re-encoding, the association of the given block is updated. A map is formed to convert the given data record encoded with the first encoding process to the second encoding process so as to preserve comparative relationships of the given data record.

Abstract: A method and system for providing a scalable storage scheme for native hierarchically structured data of relational tables, includes a base table with indicator columns with information pertaining to hierarchically structured data of a document, data tables for storing the hierarchically structured data corresponding to the indicator columns, and node identifier indexes corresponding to the data tables for mapping between the indicator columns and the hierarchically structured data in the data tables. In an embodiment, actual data for each hierarchically structured data (such as XML) column is stored in a separate data table, and each data table has a separate node identifier index. The node identifier index is searched with a key containing the document identifier and a logical node identifier is used, and a record identifier of a record in the data table containing the node assigned the logical node identifier is retrieved.

Abstract: A storage of nodes of hierarchically structured data uses logical node identifiers to reference the nodes stored within and across record data structures. A node identifier index is used to map each logical node identifier to a record identifier for the record that contains the node. When a sub-tree is stored in a separate record, a proxy node is used to represent the sub-tree in the parent record. The mapping in the node identifier index reflects the storage of the sub-tree nodes in the separate record. Since the references between the records are through logical node identifiers, there is no limitation to the moving of records across pages, as long as the indices are updated or rebuilt to maintain synchronization with the resulting data pages. This approach is highly scalable and has a much smaller storage consumption than approaches that use explicit references between nodes.

Abstract: A method, apparatus, and article of manufacture for optimizing a query in a computer system. During compilation of the query, a GROUP BY clause with one or more GROUPING SETS, ROLLUP or CUBE operations is maintained in its original form until after query rewrite. The GROUP BY clause with the GROUPING SETS, ROLLUP or CUBE operations is then translated into a plurality of levels having one or more grouping sets. After compilation of the query, a grouping sets sequence is dynamically determined for the GROUP BY clause with the GROUPING SETS, ROLLUP or CUBE operations based on intermediate grouping sets, in order to optimize the grouping sets sequence. The execution of the grouping sets sequence is optimized by selecting a smallest grouping set from a previous one of the levels as an input to a grouping set on a next one of the levels. Finally, a UNION ALL operation is performed on the grouping sets.

Abstract: A new sub-document versioning method for record storages of XML documents which uses virtual cutting points to ensure that a search tree is able to support multiple versions of sub-documents and provide efficient mechanisms for XML updating. Record boundaries and virtual cut points divide the two-dimensional space, the horizontal axis representing node identifiers in document order and vertical axis representing version numbers, into rectangles. The bottom corner of the rectangle is used to represent the rectangles and the corresponding information of the corner is added to the search tree index.

Abstract: A method for multi-versioning data of a hierarchically structured document stored in data records includes: changing document data in one or more data records, each data record assigned a record identifier, the data record including a plurality of nodes assigned a node identifier, and the document assigned a document identifier; storing an update timestamp in a base table row referencing the document identifier; storing in each changed data record a start timestamp for a start of a validity period for the changed data record and an end timestamp for an end of the validity period; and storing the start timestamp and the end timestamp in one or more node identifier index entries referencing the document identifier, the record identifier, and the node identifier. A version of the document may be obtained using node identifier index entries satisfying a version timestamp.

Abstract: A system and method for concurrency control of hierarchically structured data is provided. Lock requests on a target node are processed by exploiting ancestor-descendant information encoded into prefix encoded node identifiers (IDs). A set of implicit locks on ancestor nodes along a path from an immediate parent of a target node to a root node is derived from an explicit lock request on a target node. A logical lock tree describing existing lock modes for ancestor nodes is consulted to determine compatibility with the derived set of implicit locks. If existing lock modes for ancestor nodes are compatible with the derived set of implicit locks, a lock request on a target node is granted. Otherwise, the lock request is denied. A lock release request follows the reverse process; a target node in a particular transaction is released, as are subsequent locks on its ancestors made by the same transaction.

Abstract: A variable-length binary string is utilized to encode node identifiers in a tree for an XML document object model. A general prefix encoding scheme is followed; a node identifier is generated by the concatenation of encodings at each level of a tree along a path from a root node to another particular node. Arbitrary insertions are supported without change to existing node identifier encodings. In addition, the method provides for document order when unsigned binary string comparison is used to compare encoded node identifiers. In support of sub-document concurrency control, prefix encoding provides a way to derive ancestor-descendant relationships among nodes in a tree. Lastly, the encoding method provides a natural pre-order clustering sequence, also known as depth-first clustering. If a prefix is applied to an encoding with a level number, starting with zero at the root, width-first clustering will result. A mixed clustering can also be supported.

Abstract: An XML schema is compiled into an annotated automaton encoding, which includes a parsing table for structural information and annotation for type information. The representation is extended to include a mapping from schema types to states in a parsing table. To validate a fragment against a schema type, it is necessary simply to determine the state corresponding to the schema type, and start the validation process from that state. When the process returns to the state, fragment validation has reached successful completion. This approach is more efficient than a general tree representation. Only the data representation of the schema information is handled, making it much easier than manipulating validation parser code generated by a parser generator. In addition, only one representation is needed for schema information for both document and fragment validation. This approach also provides a basis for incremental validation after update.

Abstract: A method for conversion between a decimal floating-point number and an order-preserving format has been disclosed. The method encodes numbers in the decimal floating-point format into a format which preserves value ordering. This encoding allows for fast and direct string comparison of two values. Such an encoding provides normalized representations for decimal floating-point numbers and supports type-insensitive comparisons. Type-insensitive comparisons are often used in database management systems, where the data type is not specified for values to compare. In addition, the original decimal floating-point format can be recovered from the order-preserving format.

Abstract: Disclosed is method for processing an aggregate function. Rows that contain a reference to intermediate result structures are grouped to form groups. For each group, aggregate element structures are formed from the intermediate result structures and, if the aggregate function specifies ordering, the aggregate element structures are sorted based on a sort key.

Abstract: A new sub-document versioning method for record storages of XML documents which uses virtual cutting points to ensure that a search tree is able to support multiple versions of sub-documents and provide efficient mechanisms for XML updating. Record boundaries and virtual cut points divide the two-dimensional space, the horizontal axis representing node identifiers in document order and vertical axis representing version numbers, into rectangles. The bottom corner of the rectangle is used to represent the rectangles and the corresponding information of the corner is added to the search tree index.