Create/Edit Cube: Build Specifications

A build specification consists of one or more ordered steps that prepare a cube for querying, such as loading detail data and generating aggregate data.

Each cube has a default build specifications named LOAD_AND_AGGREGATE and SYS_DEFAULT.

Build Specifications

Lists the build specifications defined for the cube.

Click the Add icon to add a new build specification.

Click the Delete icon to delete the selected build specification.

To change a build specification, select it and change the values in the Build Specification Commands.

Name: The name of the build specification. Select the one you want to modify.

Default: Identifies the script as the default for the cube. The initial default script loads data into all the measures and aggregates it.

To make a build specification the default, click the Default column of the build specification.

Build Specification Commands

Either double-click or drag-and-drop a build command onto the build specification. Most commands display a template that provides a choice of options. You can use the templates to customize the command:

Analyze
Clear
Load
OLAP DML
PL/SQL
Solve
Group

Expression Editor

Displays the selected build specification. You can enter code directly into the editor or use the templates provided by the Build Commands list.

A new build specification begins with this structure:

BUILD SPEC name(
  LOAD
)

The expression editor provides the following tools:

Icon	Description
	Reverses the last change (undo).
	Reapplies the last change after it is reversed (redo).
	Copies the selected text to the clipboard and deletes it from the text area (cut).
	Copies the selected text to the clipboard (copy)
	Copies the contents of the clipboard to the text area (paste).
	Erases the entire contents of the text area.
	Merges the entire contents of the text area into a single line. All formatting is discarded.
	Checks the syntax of the selected text.

Syntax: Dynamically displays Valid or Invalid for the syntax shown in the text area.

Analyze

Generates optimizer statistics. These statistics are used for queries against joined cubes or a cube joined to a table or view, but not queries against a single cube.

Template

None

Example

Build Syntax: ANALYZE

Clear

Prepares the cube for a data refresh by setting all or some of the values of a cube to null (NA).

Clearing all leaf values means that all facts must be reloaded, and the aggregates for any new or changed facts must be computed. However, the aggregates are not recomputed for leaf values that stay the same.

Clearing all aggregates means that all aggregates must all be recomputed.

Template

Clear {all | leaf | aggregate} values of the cube using {parallel processes | a single process} for each cube partition.

Example

Template: Clear leaf values of the cube using parallel processes for each cube partition.

Build Syntax: CLEAR LEAVES PARALLEL

Load

Refreshes cubes and dimensions from the source data, using either parallel or serial processes.

Template

Load cube data using {parallel processes for each cube | a single process for each cube partition | only partitions with data} for all measures and all the rows from the tables.

Example

Template: Load cube using only partitions with data for all measures and all the rows from the tables.

Build Syntax: LOAD PRUNE

OLAP DML

Executes an OLAP DML command or program, using either parallel or serial processes. This command provides the basic syntax for executing DML commands. You must enter the OLAP DML between the single quotes in the EXECUTE clause.

Template

Use {parallel processes | a single process} for each cube partition for all measures and for all dimension members to execute the OLAP DML.

Example

Template: Use parallel processes for each cube partition for all measures and for all dimension members to execute the OLAP DML.

Build Syntax: EXECUTE OLAP DML 'MYFORECASTPRG' PARALLEL

PL/SQL

Executes a PL/SQL procedure or script. This command provides the basic syntax for executing PL/SQL commands during a build. You must enter the PL/SQL between the single quotes in the EXECUTE clause.

Template

None

Example

Build Syntax: EXECUTE PLSQL ' DBMS_CUBE.BUILD(q'!GLOBAL."TIME", GLOBAL.CHANNEL, GLOBAL.CUSTOMER, GLOBAL.PRODUCT, GLOBAL.UNITS_CUBE USING (LOAD NO SYNCH WHERE UNITS_FACT.MONTH_ID LIKE '2006%' AND UNITS_FACT.SALES > 5000)!')'

Solve

Aggregates the measures of the cube, using either parallel or serial processes.

Template

Solve cube using {parallel processes | a single process} for each partition for all measures and all dimension members.

Example

Template: Solve cube using a single process for each cube partition for all measures and all dimension members.

Build Syntax: SOLVE SERIAL

Group

Runs one or more commands over a group of measures instead of all the measures in the cube. This command provides the basic structure of a FOR loop in PL/SQL for you to edit.

You can select all dimension members, dimension members by level, or individual dimension members. To select all members, select select all members). To select members by level, select select all members at the level(s), then from the Levels dialog box, select the levels to include.

To include or exclude selected dimension members, select include members or exclude members. Click the dimension member displayed. From the Select Dimension Member dialog box, select one or more dimension members to include or exclude.

You can also select the measures. Click the list of measures in the template to display the Measure Selection dialog box. All measures in the cube are in the initial list.

Template

For dimension dimension_name {select all members | select all members at the level(s) levels_list | include members dimension member | exclude members dimension member}.

Run the following command(s) for these measures: list of measures

Example

Template:

For dimension TIME select all members at level(s) MONTH.

For dimension PRODUCT select all members.

Run the following command(s) for these measures: UNIT_COST, UNIT_PRICE.

Build Syntax:

FOR
    "TIME" LEVELS ("TIME"."MONTH"), 
      PRODUCT ALL
    MEASURES
    (
    PRICE_CUBE.UNIT_COST, PRICE_CUBE.UNIT_PRICE)
  BUILD
  (
  LOAD
  )