Glosario de Bases de Datos Sybase, "P,Q"

A 2K block of data that is the minimal unit that can be read from or written to disk.

See partition.

A page lock locks an entire data or index page. Adaptive Server uses page locks as frequently as possible to reduce the contention for data among users and to improve concurrency. A page lock is less restrictive than a table lock.

An action performed by Adaptive Server when new data or index rows need to be added to a page , and there is not enough room for the new rows. Usually, the data on the existing page is split approximately evenly between the newly allocated page and the existing page.

The allocation of a new last page for a partition from a device or extent that was not originally assigned to the partition. This action is performed as needed by Adaptive Server.

A partition-based access method in which multiple worker processes simultaneously scan data pages in a partitioned table when the clustered index key matches a search argument. One worker process is assigned to each partition. A worker process may also traverse the clustered index first to find the first data page in the partition that satisfies the search argument.

A hash-based access method in which multiple worker processes scan a single chain of data pages in a table at the same time and apply an internal hash function to each page ID. The hash function determines which worker process reads the rows in current page. Typically, the number of worker processes is no more than two or three because it takes only two to three worker processes to fully utilize the I/O of a given physical device.

A hash-based access method in which multiple worker processes simultaneously scan level 0 index pages and apply a hash function to the data page IDs or the key values in each index page. The hash function determines which worker process scans the data rows referenced by the data page ID or key value.

A partition-based access method in which multiple worker processes simultaneously scan partitions in a partitioned table. The number of worker processes is equal to the number of partitions.

The simultaneous execution of Adaptive Server tasks or subtasks. Multiple processes share memory or use some type of synchronized method for passing messages between them.

A technique that employs multiple worker processes to sort data in parallel, either in response to a create index command or to a query that requires an internal sort. A single process partitions the input data into discrete ranges; multiple processes simulataneously sort each individual range of data rows and create subindexes; then a single process merges the sorted ranges and indexes into one sorted index or result set.

An argument to a stored procedure or system procedure .

A linked chain of database pages that stores a database object . You can divide a table into multiple partitions that reside on separate physical devices to improve the performance of concurrent inserts and to facilitate parallel queries and sorts.

A parallel access method that uses two or more worker processes to access separate partitions of a table. This category includes parallel partition scan and parallel clustered index partition scan .

A mode that allows clients to communicate with remote servers in native syntax. The Transact-SQL parser and compiler are bypassed in this mode, and each language "batch" received from the user is passed directly to the server to which the user is connected.

The speed with which Adaptive Server processes queries and returns results. Performance is affected by several factors, including indexes on tables, use of raw partitions compared to files, and segments.

The authority to perform certain actions on certain database objects or to run certain commands.

Occur when one transaction reads a set of rows that satisfy a search condition, and then a second transaction modifies the data (through an insert , delete, update , and so on). If the first transaction repeats the read with the same search conditions, it obtains a different set of rows.

Mapping the logical design to the Transact-SQL data definition commands that actually create the databases on the server.

A column name, or set of column names, used in a create index statement to define an index on a table. A physical key on a table is not necessarily the same as the logical key .

A disk I/O to access a data, index, or log page. Adaptive Server estimates physical reads and logical reads when optimizing queries. See also logical read .

A query that restricts results to a single specific value, usually using the form "where column_value = search_argument".

The maximum number of decimal digits that can be stored by numeric and decimal datatypes. The precision includes all digits, both to the right and to the left of the decimal point.

A clause that is based on a comparison operator (=, >, <, <=, >=, != ,<> , !> , or !<). It does not apply to and, or , or not .

The process of performing multipage I/O's on a table, nonclustered index, or the transaction log. For logs, the server can fetch up to 256 pages, for nonlog tables and indexes, the server can fetch up to eight pages.

Used to refer to keys in a composite index. Search values form a prefix subset when leading columns of the index are specified. For an index on columns A, B, and C, these are prefix subsets: A, AB, ABC. These are not: AC, B, BC, C. See also matching index scan and non-matching index scan .

The column or columns whose values uniquely identify a row in a table.

A primary key constraint is a unique constraint that does not permit null values for the component key columns. There can only be one primary key constraint per table. The primary key constraint creates a unique index on the specified columns to enforce this data integrity.

See permissions .

An execution environment scheduled onto physical CPUs by the operating system.

A process in which a certain Adaptive Server task runs only on a certain engine or a certain engine runs only on a certain CPU.

In a parallel sort , producer processes read data simultaneously from the input table, determine the range to which each data value belongs, and distribute data values to consumer processes associated with the proper ranges.

A subset of the columns in a table. It is one of the basic query operations in a relational system. See also selection and view .

The ability to impersonate another user in the server. A System Security Officer can grant proxy authorization to a user. Proxy authorization allows administrators to check permissions for a particular user or to perform maintenance on a user's database objects. Application servers can log into the server and execute procedures and commands on behalf of several users.

A local table that has been mapped to a table on a remote server. The proxy table contains metadata and is used to access the remote table as if it were a local table.

The name of a database object can be qualified, or preceded by, the name of the database and the object owner.

A SQL statement or group of SQL statements that access and/or manipulate data in a database. See also data retrieval .

The ordered set of steps required to carry out a query , complete with the access methods chosen for each table.

An internal tree structure that represents the user's query . A large portion of query processing and compilation is built around the shape and structure of this internal data structure. For stored procedures, views, triggers, rules and defaults these tree structures are stored in the sysprocedures table on disk, and read back from disk when the procedure or view is executed. See also source text .