SQL/MX 3.2 Management Manual (H06.25+, J06.14+)

Resource forks contain similar information to Guardian file labels, but have the capacity to store
a much greater variety and quantity of information. Examples of data fork/resource fork pairs are:
• $DATA.ZSDADMM8.WDE6TI00 / $DATA.ZSDADMM8.WDE6TI01
• $DATA.ZSDADMM8.HEPQTI00 / $DATA.ZSDADMM8.HEPQTI01
• $DATA.ZSDADMM8.SS8BSI00 / $DATA.ZSDADMM8.SS8BSI01
Data files are sometimes referred to as data forks to distinguish them from resource forks. Under
normal circumstances, you are not expected or permitted to refer explicitly to resource forks by
name. The one significant exception is TMF recovery using Guardian file names.
Designating Wild Cards in SQL/MX Schema Names
An SQL/MX schema has a designated subvolume in which every partition belonging to that schema
must reside. Therefore, in general, a Guardian wild card of the form $*.subvolume.* can serve
as convenient shorthand for every Guardian file in the SQL/MX schemas whose schema subvolume
is subvolume.
NOTE: If you do not specify a subvolume name for a schema’s partitions when you create a
schema, NonStop SQL/MX provides a unique subvolume name by default. In doing this, NonStop
SQL/MX establishes and maintains a one-to-one correspondence between schemas and subvolumes.
All files in a schema subvolume are SQL/MX files. The DP2 process does not allow other types of
files to be created in a schema subvolume.
For more information, see the SQL/MX Reference Manual.
Using the LOCATION Clause
Both CREATE TABLE and CREATE INDEX include a LOCATION clause for specifying the physical
locations of partitions. The complete syntax for this LOCATION clause is:
LOCATION [\node.]$volume[.subvolume.file-name] [NAME partition-name]
where subvolume is the name of the designated schema subvolume for the schema in which the
table or index is being created, and file-name is a Guardian file name that conforms to the
rules specified in “SQL/MX Subvolume Naming Guidelines” (page 73). Otherwise, an error is
generated.
When you create SQL/MX tables and indexes, you would typically use the LOCATION clause to
partition data on particular volumes, without regard to the Guardian subvolume and file name
assigned to these partitions. However, it might be necessary to re-create a file with the exact same
Guardian names that it was given when first created.
When you use a subvolume in the LOCATION clause, it must be the one that is defined for the
schema.
For more information, see the SQL/MX Reference Manual.
Using the Subvolume Option in the CREATE SCHEMA Statement
The CREATE SCHEMA statement accepts an optional subvolume name, a nonstandard HP extension.
This option enables you to create an RDF backup database. The subvolume name of each backup
schema must match the subvolume name of the corresponding primary schema. If you are not
creating an RDF backup of the database, you should omit the optional subvolume, in which case
NonStop SQL/MX generates a new and unique subvolume name for the schema. In all cases, the
schema subvolume name is written to the SCHEMA_SUBVOLUME column of the system schema
table NONSTOP_SQLMX_system-name.SYSTEM_SCHEMA.SCHEMATA.
The format of the CREATE SCHEMA statement is:
CREATE SCHEMA schema-clause [schema-element]
74 Creating an SQL/MX Database