E Deleting Unwanted Objects from the Repository

If your repository contains many objects that you do not need, you can use the command-line pruning utility to delete the unwanted objects. This appendix explains how to the pruning utility works and how to use it.

This appendix contains the following topics:

• About the Object Pruning Utility

• Using the Object Pruning Utility

• Deletion Rules for the Object Pruning Utility

About the Object Pruning Utility

If you have a large number of extraneous or unwanted objects in your repository, you can delete the unwanted objects using the prunerpd command-line utility. You can use prunerpd on both Windows and UNIX systems.

You can delete unwanted repository objects such as databases, tables, columns, initialization blocks, and variables. However, note that the pruning utility does not remove objects from the Oracle BI Presentation Catalog.

Deleting objects from the repository has a cascading effect. For example, if a physical column is deleted, then any mapped logical columns are deleted, as well as any associated presentation columns. See "Deletion Rules for the Object Pruning Utility" for more information.

Using the Object Pruning Utility

You must first create the input file that contains the list of repository objects to be deleted. Then, you must run the utility at the command line, passing the input file as an argument

This section contains the following topics:

• Creating the Input File

• Running the Prune Utility

Creating the Input File

The prune utility accepts the list of repository objects you want to delete as a text file. The utility can accept multiple input files at a time. The syntax rules for the input file are shown in Table E-1.

Note: Object names in the input file must match the fully qualified name that is used in the repository. Wildcards (such as "*" and "?") are not supported in the object name.

Table E-1 Syntax Rules for Input File

Object Type	Example	Action
Database	D "Paint"	Deletes the database named "Paint."
Table	T "W_AGREE_D"	Deletes the table or alias named "W_AGREE_D" from the Physical layer.
	T "DB"."Catalog"."Schema"."Table"	Deletes the table or alias named "Table" from the schema named "Schema," contained in the catalog named "Catalog," located in the database named "DB," from the Physical layer.
Column	C "W_AGREE_MD"."AGREE_CD"	Deletes the column named "AGREE_CD" located in a table or alias named "W_AGREE_D" from the Physical layer.
Initialization block	I "External Metadata Strings"	Deletes the initialization block named "External Metadata Strings."
Variable	V CURR_USER	Deletes the variable named "CURR_USER."

For example, a text file that contains instructions to delete a database named "Stock Quotes" and a physical column named "S_NQ_ACCT"."USER_NAME" would include the following entry:

D "Stock Quotes" C "S_NQ_ACCT"."USER_NAME"

Use white space as a a delimiter in the input file (a single space, tab, or multiple spaces).

Running the Prune Utility

Before running prunerpd, you must first run bi-init.cmd (or bi-init.sh on UNIX systems) to launch a command prompt that is initialized to your Oracle instance. You can find this utility in:

ORACLE_INSTANCE/bifoundation/OracleBIApplication/coreapplication/setup

Then, run prunerpd from the resulting command prompt window with the desired options. The prune utility accepts the following parameters:

-s

Provide the location of the source repository file.

-p

Provide the repository password.

The password argument is optional. If you do not provide a password argument, you are prompted to enter a password when you run the command. To minimize the risk of security breaches, Oracle recommends that you do not provide a password argument either on the command line or in scripts. Note that the password argument is supported for backward compatibility only, and will be removed in a future release.

-f

Provide the input file name (in text format) that contains the list of repository objects to be removed. Separate multiple file names by spaces. Enclose spaces within a filename with double quotes (" ").

-o

Provide the name and location of the output repository file, also known as the pruned repository.

-l

Provide the name and location of the output log file. All actions performed on the repository are written to this file, including descriptions.

-e

Provide the name and location of the error log file. The pruning utility writes exceptions and errors to this log.

-8

Input is in UTF-8 format.

Example E-1 Prune Utility Usage Example

prunerpd -s C:/OBI/Server/Repository/BIApps.rpd 
-f "C:/Remove Oracle EBS Objects.txt"
-o "C:/OBI/Server/Repository/BIApps Pruned.rpd"
-l "C:/temp/BIApps Prunning.log" -e "C:/temp/ BIApps Prunning.err"
Give password: my_repos_password

Log File

The log file includes a list of actions performed on the repository file and is in XML format. Other messages, such as progress indicators, are sent to the standard output stream.

Error Log File

The error log file includes a list of errors encountered while processing the contents of the input file and is in XML format. Other errors are sent to the standard output error stream.

Deletion Rules for the Object Pruning Utility

Deleting repository objects has a cascading effect. This section describes the deletion rules.

Physical Layer Rules

• If a physical column or a table is deleted, then all of the affected keys, foreign keys, and complex joins are deleted as well. The internal obsolete attribute definition (attr defn) that links a logical column to a physical column is also removed.

• Empty schemas, catalogs, and databases are removed.

• If a table is deleted, then all its columns are deleted.

Logical Table Rules

• If a regular column (not an aggregate or derived column) is not mapped in any logical source, then it is deleted. The keys, including the level key and the logical key, are also removed.

• If the source column for a derived column or its referenced variable is deleted (corrupted), then the column is removed.

• If an aggregate rule or override aggregate rule for an aggregate column is corrupted (due to a logical column deletion), then the column is removed.

• If a logical table is removed (because its underlying physical table was deleted), then the keys, foreign keys, logical joins, sources, and source folder are removed.

• If a logical table source does not have any valid mapping, then it is deleted.

• If a logical table source is retained, but its aggregate content or filters are corrupted, then the corresponding expressions are set to null. The join specification is also removed.

• If a logical table, dimension, or business model is empty (contains no meaningful child), then it is deleted.

Presentation Layer Rules

• If a logical column is removed (because its underlying physical column was deleted), then any corresponding presentation columns are removed.

• If a presentation table or subject area does not contain children, then it is removed.

Security Rules

• If a security filter for a user or application role becomes corrupt due to deletion, then the filter is removed. If all filters are removed for a user or application role, then the internal privilege object is deleted.

• Even if all filters for an application role are deleted, the application role is still maintained.

• To remove an application role from the repository, you must explicitly delete it. See Oracle Fusion Middleware Security Guide for Oracle Business Intelligence Enterprise Edition for information about deleting application roles.

Variable Rules

• Initialization blocks are deleted if the underlying connection pool is deleted.

• Repository and session variables are deleted if the associated initialization blocks are deleted.

• If a session variable is deleted and its parent initialization block does not contain variables, then the initialization block is removed.

• If an initialization block is deleted, then its variables are removed.

Marketing Rules

• Qualified list items are deleted if the associated cache catalog, GUID column, or qualified column is deleted.

• Target levels are deleted if the associated catalog (Segmentation Catalog name) is deleted.

• List catalogs are deleted if the associated catalog, table, or column is deleted.

• Conforming dimensions are deleted if the associated catalog, table, or column is deleted.