SQL/MX Glossary

Glossary
HP NonStop SQL/MX Glossary—523732-002
Glossary-30
repeatable read
repeatable read. The level of consistency that repeatable read access provides. Contrast
with cursor stability. See also repeatable read access.
repeatable read access. An access option for transaction consistency that ensures that the
data accessed by a transaction cannot be changed by other users until the transaction
ends. With repeatable read access, a process tests for existing locks on data before
acquiring its own lock and holds the lock until the transaction completes. The
implementation for repeatable read access is equivalent to serializable access.
Contrast with read uncommitted access and read committed access.
reserved word. A word reserved for use by SQL/MX. A reserved word cannot be used as a
regular identifier.
resource fork. One of two physical files instantiated when an SQL/MX object is created.
One file is the data fork, where user data resides, and is similar to an SQL/MP format
object. The other, the resource fork, is a new file that contains more structural
descriptions of the table. In contrast, SQL/MP keeps all its structure information in its
metadata and DAM labels. Because DAM has a limitation on how much label
information can be stored, SQL/MX keeps all its run-time metadata information, such
as the partition array, in the resource fork. Resource forks are created for all persistent
SQL/MX objects, such as tables, SQL/MX temp tables, SQL/MX worktables, and
indexes.
result table. The output that results from the execution of a SELECT statement.
right join on. A type of join that joins rows that satisfy the search condition in the ON clause
of a join operation, plus rows from the right table that do not satisfy the search
condition.
routine. A set of program statements that performs a specific operation. A routine can
accept arguments and return values. Functions and procedures are types of routines.
See also function, procedure, and user-defined routine (UDR).
row. A horizontal component of a table; the relational representation of a record. A row
contains one data value for each column in the table. Contrast with column
.
row lock. A lock held by a process on one specific row. Contrast with partition lock
and
table lock.
row value constructor. A series of expressions enclosed in parentheses or a row subquery
(subquery returning a single row). Row value constructors are used as operands of
predicates.
rowset. A group of rows retrieved by an SQL cursor in a single FETCH operation.
run unit. A set of one or more object programs in the same object program file.