How can I access a registered table in a remote Oracle database?

Create an Oracle database link. To access remote tables in Oracle, a link must be defined to the remote node. The Oracle 'create database link' statement performs this operation:


CREATE [PUBLIC] DATABASE LINK
CONNECT TO
IDENTIFIED BY
USING ''

The linkname becomes the name by which tables at the remote database can be identified.

The remote database must be accessible to the local database via SQL*Net.

The username can refer to any valid user on the remote database (i.e., it does not necessarily have to be the Documentum dbo).

Once the link is created, tables on the remote database may be accessed by the following tablename syntax:
[owner.]tablename@linkname
Accessing remote Oracle tables through Documentum:

DQL does not support the @linkname syntax. In order to access remote
tables, the following steps are required:

1) Create the database link. This may be done either from Documentum (via the exec_sql function of the Apply method), or directly in Oracle. It can be either a public or private database link. You must have DBA privilege to create a public link. If it is a private link, it must be created by the Documentum dbo.

2) Create synonyms in the local database for the tables in the remote database. These MUST be private synonyms, they must be created by the

Documentum dbo and you need the Oracle CREATE SYNONYM privilege to create these.
3) Use the register table statement to register the synonyms.

DQL Queries - Help

Help on DQL Queries:


1. Creating a user
create “dm_user” object
set client_capability=2,
set default_folder=’’,
set home_docbase=’’,
set user_address=’a@abc.com’,
set user_os_domain=’’,
set user_name=’’,
set user_os_name=’’,
set user_privileges=0;

2. DQL to execute stored procedure (works for all supported DBMSes)
execute exec_sql with query=’execute my_sp_test ”123”,”prasad” ‘;

3. Running DQL in batch mode
a. Create a dql file
b. Run the following from command prompt
$idql docbasename -Uusername -Ppassword -Rtest.dql
Example command
C:\Documentum\product\5.3\bin>idql32 docbasename-Uusername -Ppassword -RC:
\test.dql
test.dql would contain something like this

go

go

4. DQL to Get all Files under a Particular Cabinet
select r_object_id, object_name from dm_document(all) where folder(’/Cabinet name’, descend);
The above DQL gives all versions. To get only current versions
select * from dm_document where folder (’/Cabinet name’, descend)

5. DQL to get total number of documents and folders under a cabinet
SELECT count(*) as cnt, ‘Docs’ as category FROM dm_document(all)
WHERE FOLDER (’/Cabinet Name’,DESCEND)
UNION
SELECT count(*) as cnt, ‘Folders’ as category FROM dm_folder
WHERE FOLDER (’/Cabinet Name’,DESCEND)

6. DQL to find whether a document is a part of virtual document
SELECT object_name,r_object_id FROM dm_sysobject
WHERE r_object_id IN
(SELECT parent_id FROM dmr_containment
WHERE component_id = (SELECT i_chronicle_id FROM dm_sysobject WHERE r_object_id = ‘’))

7. Repeating attributes
Repeated attribute queries are always a fun. I am going to present here more repeated attribute queries. But for starters, recognize the importance of ANY keyword
select r_folder_path from dm_folder where object_name=’myFolder’ and any r_folder_path =’/Cabinet Name/test’;

8. DQL to find object type of a document

select r_object_type from dm_document where object_name=’ObjectName’;

9. DQL for index

Following query creates index

EXECUTE make_index WITH type_name=’dmi_workitem’,attribute=’r_workflow_id’
Using the query below, one can find out whether index has been succesfully created or alternatively whether the index exists or not

Select r_object_id,index_type,attribute,attr_count,data_space from dmi_index where index_type in (select r_object_id from dm_type where name=’dmi_workitem’);

10. DQL to see sessions
execute show_sessions

11. Enable FTDQL
SELECT r_object_id, score, text, object_name, r_object_type, r_lock_owner, owner_name, r_link_cnt, r_is_virtual_doc, r_content_size, a_content_type, i_is_reference, r_assembled_from_id, has_frzn_assembly,
a_compound_architecture, i_is_replica, r_policy_id,acl_name, r_creation_date, r_modify_date, subject
FROM custom_document WHERE (custom_attr1 = ’search1′ AND (custom_attr2 = ’search2′ AND
)) AND (a_is_hidden = FALSE) ENABLE (FTDQL)

12. DQL to get current date, time

select DATE(now) as systime from dm_server_config;

13. DQL to list all available templates in Webpublisher
select * from my_document where folder(’/WebPublisher Configuration/Content Templates/myTemplates’,descend) and any r_version_label =’Approved’;

14. DQL to list objects having duplicate names
SELECT object_name, count(*) FROM dm_document

GROUP BY object_name

HAVING count(*) > 1

ORDER BY object_name


15. Clear INBOX



delete dmi_queue_item objects where delete_flag=0



16. DQL to retrieve all required attributes of a particular type

SELECT attr_name FROM dmi_dd_attr_info WHERE type_name=’dm_document’ AND is_required <> 0



17. DQL to list workflow attachments



select r_component_id, r_component_name from dmi_wf_attachment where r_workflow_id = ‘’



18. If your statistics are not up to date, database may choose a very inefficient execution plan. Be sure toupdate statistics often. It is recommended that you use the dm_UpdateStatistics job as it will calculateextended statistics on particular tables and columns which provide additional performance benefits. If the DBA uses their own scripts to calculate the statistics, then these enhancements will not be available.



19.Use the script “dctm_indexes_by_table.sql” to generate a list of all indexes on Documentum tables, ordered by table name. (Available from http://developer.documentum.com.)

Use the script “dctm_indexes_by_index.sql” to generate a list of all indexes on Documentum tables,

ordered by index name. (Available from http://developer.documentum.com.)



20.It is strongly recommended that all indexes on Documentum base tables be created from within Documentum and not at the database level.

There are two reasons for this:

The internal conversion process from DQL to SQL will check for the presence of a dmi_index object for repeating valued attributes and will generate different SQL according to what it finds.

If indexes are created directly through SQL*Plus, then Documentum will not know anything about them and will assume it is unindexed. This may result in a less efficient SQL query.

Secondly, if the indexes are created from within Documentum, and they are inadvertently dropped, the dm_DBWarning job will automatically recreate them at next execution.

The syntax for creating new indexes is as follows:



In DQL:

EXECUTE make_index WITH type_name=object_type,

attribute=attribute_name{,attribute=attribute_name,…)

or

Using APIs:

dmAPIGet(”apply,session,NULL,MAKE_INDEX,TYPE_NAME,S,object_type,ATTRIBUTE,S

,attribute_name(,ATTRIBUTE,S,attribute_name,…}

These indexes will be created in the tablespace identified by the index_store server.ini parameter.

Be sure to create the index on the appropriate type table.

For example, although ‘keywords’ is an attribute of the dm_document type, it is actually inherited from ‘keywords’ from the dm_sysobject type.



Drop the Index if Necessary

If there is no performance improvement, OR the optimizer is not using the new index, drop it using:

EXECUTE drop_index [[FOR] dmi_index_id] [WITH name = index_name]

or

dmAPIGet(”apply,session,dmi_index_id,DROP_INDEX [,NAME,S,index_name]“)



21. Using the iapi utility and the trace API you can generate log information that contains the SQL resulting from a DQL query.

This is useful when it is a DQL query that is performing poorly and you wish to test using a nonsuperuser account. Non-query type APIs are not traced.

Here is an example of tracing a simple DQL query.

trace,c,10,,DM_QUERY.



More on TKPROF here



http://download-west.oracle.com/docs/cd/B10501_01/server.920/a96533/sqltrace.htm



22. DQL Hints



SELECT object_name FROM dm_document ENABLE (RETURN_TOP 10)



For DB2, performance can be improved using

SELECT object_name FROM dm_document ENABLE (RETURN_TOP 10, OPTIMIZE_TOP 10)



The FETCH_ALL_RESULTS N hint fetches all the results from the database

immediately and closes the cursor. The hint does not affect the execution plan,

but may free up database resources more quickly.

To fetch all the results, set N to 0.

On SQL Server, it is recommended that you use SQL_DEF_RESULT_SETS

instead of the FETCH_ALL_RESULTS hint. SQL_DEF_RESULTS_SETS

provides the same benefits and is the recommended way to access SQL Server

databases.



Passthrough hints are hints that are passed to the RDBMS server. They are not

handled by Content Server.



To include a passthrough hint, you must identify the database for which the

hint is valid. To identify the target database, keywords precede the hints. The

valid keywords are: ORACLE, SQL_SERVER, SYBASE, and DB2. For

example, the following statement includes passthrough hints for SQL Server:

SELECT “r_object_id” FROM “dm_document”

WHERE “object_name” =’test’

ENABLE SQL_SERVER(’ROBUST PLAN’,'FAST 4’,’ROBUST PLAN’)

For portability, you can include passthrough hints for multiple databases in

one statement. The entire list of hints must be enclosed in parentheses. The

syntax is:

(database_hint_list {,database_hint_list})

where database_hint_list is:

db_keyword(’hint’{,’hint})

db_keyword is one of the valid keywords identifying a database. hint is any hint

valid on the specified database.

For example:

SELECT object_name FROM dm_document doc dm_user u

WHERE doc.r_creator_name = u.user_name ENABLE

(ORACLE(’RULE’,’PARALLEL’), SYBASE(’AT ISOLATION READ

UNCOMMITTED’),SQL_SERVER(’LOOP JOIN’,’FAST 1’)



Use FETCH_ALL_RESULTS if you want to reduce the resources used by the

database server by quickly closing cursors. On SQL Server, try

FETCH_ALL_RESULTS if using SQL_DEF_RESULT_SETS did not improve

query performance.



23. If the table is registered, check with the owner of the registered table, or a Documentum superuser, about giving you access to the table.



You will need at least a BROWSE permit on the registered table object in order to access it. If the table has not yet been registered, check with the table’s owner about registering it. Note that you must have READ access to the dm_registered object for this table in order to access it in any way (SELECT, INSERT, UPDATE, DELETE).



Note that if you specified the special dm_dbo document base owner, the first parameter will hold the actual name of the document base owner.



24. DQL to list docbrokers



execute list_targets



25. DQL to list worrkflow information



select task_name, task_state, actual_start_date, dequeued_date from dmi_queue_item where router_id = ‘workflowId’



26. Query to find active workflows, supervisors of the workflows



select r_object_id, object_name, title, owner_name,r_object_type, r_creation_date, r_modify_date, a_content_type from dm_document where r_object_id in(select r_component_id from dmi_package where r_workflow_id in (select r_object_id from dm_workflow where r_runtime_state = 1))



26. Though this is not DQL, I thought of mentioning here as it most widely used DFC operation



sysObject.queue (”dm_autorender_win31″, _

“rendition”, _

0, _

False, _

dueDate, _

“rendition_req_ps_pdf”)



27. To find ACLs related to dm_sysobject:

select r_object_id as obj_id, object_name from



dm_sysobject (all)



where acl_name =” and



acl_domain = ”

To find ACLs as the default ACL of a user:

select user_name from dm_user where acl_name=”

To find ACLs associated with a type:

select r_object_id, r_type_name from dmi_type_info where acl_domain=” and acl_name=”

When these queries do not return any related objects and you still cannot delete the ACL, use tracing:

In the Message Tester or IAPI, execute the command: apply,c,NULL,SQL_TRACE,LEVEL,I,1

Then, try to delete the ACL that is causing problem.

Turn off tracing by executing:

apply,c,NULL,SQL_TRACE,LEVEL,I,0

Examine the session log for the trace output, located in $DOCUMENTUM/dba/log//

Excerpt of the session log:

[DM_ACL_E_DESTROY_IN_USE]error: “Failed to destroy the ACL ‘Documentum Users’ in domain ‘dbabep’ because it is in use.”



[DM_SESSION_I_SESSION_QUIT]info: “Session 01000dcb80010ccf quit.”

Run the following query from Oracle SQL:

select r_object_id from dm_sysobject_s where acl_domain =’dbabep’ and acl_name =’Documentum Users’ union select r_object_id from dm_user_s where acl_domain = ‘dbabep’ and acl_name =’Documentum Users’ union select r_object_id from dmi_type_info_s where acl_domain =’dbabep’ and acl_name =’Documentum Users’

It will return the r_object_id value = 09000dcb800362c4.

Run the following query from IDQL:

select object_name



from dm_sysobject



where r_object_id=’09000dcb800362c4′

If the above query returned nothing, then this object cannot be accessed by any Documentum WorkSpace method, only via SQL in the underlying Oracle database. The following entries must be deleted at the RDBMS level: dm_sysobject_s and dm_sysobject_r tables; there are no entries in the dmi_object table



28. List of object types and corresponding identifiers. Helpful when reading the code.



00 dmi_audittrail_attrs

03 dm_type

05 dmr_containment

06 dmr_content

08 dm_application

08 dm_job

08 dm_procedure

08 dm_query

08 dm_script

08 dm_smart_list

09 dm_document

0b dm_folder

0c dm_cabinet

0d dm_assembly

10 dm_method

11 dm_user

12 dm_group

19 dm_registered

1f dmi_index

26 dmi_registry

27 dm_format

28 dm_filestore

28 dm_store

2c dm_distributedstore

2e dmi_type_info

2f dm_dump_record

30 dmi_dump_object_record

31 dm_load_record

32 dmi_load_object_record

37 dm_relation

3a dm_location

3b dm_fulltext_index

3c dm_docbase_config

3d dm_server_config

40 dm_blobstore

41 dm_note

45 dm_acl

46 dm_policy

49 dmi_package

4a dmi_workitem

4c dm_activity

4d dm_workflow

53 dm_literal_expr

5e dm_federation

5f dm_audittrail_acl

5f dm_audittrail_group

5f dm_audittrail

66 dm_alias_set

6a dmi_dd_attr_info

0b dm_taxonomy

0b dm_xml_application

6b dm_display_config

20 dmi_sequence



29. Job scheduler Query



SELECT ALL r_object_id, a_next_invocation

FROM dm_job

WHERE (

(run_now = 1)

OR ( (is_inactive = 0)

AND ( ( a_next_invocation <= DATE(’now’)

AND a_next_invocation IS NOT NULLDATE )

OR ( a_next_continuation <= DATE(’now’)

AND a_next_continuation IS NOT NULLDATE )

)

AND ( (expiration_date > DATE(’now’))

OR (expiration_date IS NULLDATE))

AND ( (max_iterations = 0)

OR (a_iterations < max_iterations) )

)

)

AND (i_is_reference = 0 OR i_is_reference is NULL)

AND (i_is_replica = 0 OR i_is_replica is NULL)

ORDER BY a_next_invocation, r_object_id



29. DCTM 5.3SP1 dmclean has a new argument -clean_aborted_wf. If specified, the method removes all aborted workflows in the repository. You can set this using DA





30. To deal with apostrophes in DQL, escape with apsotrophe. For example to find all users with name like O’Hare, use following DQL



select group_name from dm_group where any users_names like ‘% O”Hare%’



31. In order to view completed workflows, you need to enable workflow auditing. Workflow Manager has an option to turn auditing ON within a workflow template to view completed workflows.



From Workflow Manager:



1) Uninstall template

2) File > Template Properties

3) Under Template Audit Trail Setting, select option ‘Always On’ - Audit trail data for each instance will be available and saved at workflow completions.



32. In a workflow, a performer can enter comments while he/she is performing the task. And those comments are carried to the next performer in this workflow. When using WebPublisher workflow report to review the workflow instances, and in workflow history, you are only able to get the truncated comments back from WP’s interface. Especially for those already aborted workflow instances, there is no way that you can get the complete comments back from WP’s GUI. In some cases, those comments are very important to our customers and they need to find a way to get those comments back.



Here is the steps to get those comments back:



==



1) Identify the workflow from the dm_workflow table, get the r_object_id of the workflow:



select r_object_id, object_name from dm_workflow where object_name = ‘your work flow name’

2) Identify the notes that are carried by this workflow:



select r_note_id from dmi_package where r_workflow_id = ‘ the object id of the workflow’

3) Get the content id of each of those note ids returned:



select r_object_id from dmr_content where any parent_id = ‘the note id’

4) Go to DA, Administration->Job Management->Administration, use the “GET_PATH” method to find out the path of the files which stores the comments.



33. Query to get all documents expired in previous 1 month



SELECT s.r_object_id, s.object_name, DATETOSTRING(”r.a_expiration_date”,’mm/dd/yyyy’) as creation_date

FROM dm_sysobject_s s, dm_sysobject_r r

WHERE s.r_object_id = r.r_object_id

AND s.r_object_type = ‘dm_document’

AND DATEDIFF(month,”r.a_expiration_date”,DATE(NOW)) >= 0

AND DATEDIFF(month,”r.a_expiration_date”,DATE(NOW)) <= 1

AND r.r_version_label = ‘Expired’

ORDER BY 3



34. DQL to find all the folders in a Cabinet, where the folders are contentless

SELECT f1.object_name, f1.r_object_id, f1.r_folder_path

FROM dm_folder f1

WHERE FOLDER(’/Cabinetname’,descend)

AND NOT EXISTS (SELECT f2.object_name FROM dm_sysobject f2 WHERE ANY f2.i_folder_id = f1.r_object_id)

ORDER BY object_name



35. DQL for finding all checked out documents in a docbase

select * from dm_document where (r_lock_owner is not nullstring or r_lock_owner <> ” or r_lock_owner <> ‘ ‘)



36.DQL to list the users who has access to particular folder path

SELECT i_all_users_names FROM dm_group

WHERE group_name IN (SELECT r_accessor_name FROM dm_acl

WHERE object_name IN (SELECT acl_name FROM dm_folder

WHERE ANY r_folder_path = ‘/folderpath’))

ORDER BY i_all_users_names



37. Query to find out what user signed off on what document



SELECT “audited_obj_id” FROM “dm_audittrail” WHERE

“event_name” = ‘dm_signoff’ AND

“user_name” = ‘tom’ AND

substr (”audited_obj_id”, 1, 2) = ‘09′AND

“time_stamp” >= DATE(’01/01/1998′, ‘dd/mm/yy’) AND

“time_stamp” <= DATE(TODAY)



38. Is there any way or method by which we can can get to know about the period , whenever a metadata(attribute) is added to object or any of its sub object types.

You would need to register the save event to be audited on your object type. Whenever a property is changed it will have a save event to write it to the content server.



39. listing all contents that are in WIP state only. If a content is also in Staging or Active or Expired state, it will not show up in the result:
DQL> SELECT object_name, r_version_label FROM dm_sysobject where any r_version_label in (’WIP’) and r_object_id not in (select r_object_id from dm_sysobject where any r_version_label in (’Staging’, ‘Approved’, ‘Expired’) )

39. Query to get all expired documents in previous month

SELECT s.r_object_id, s.object_name, DATETOSTRING(”r.a_expiration_date”,’mm/dd/yyyy’) as creation_date FROM dm_sysobject_s s, dm_sysobject_r r WHERE s.r_object_id = r.r_object_id AND s.r_object_type = ‘dm_document’ AND DATEDIFF(month,”r.a_expiration_date”,DATE(NOW)) >= 0
AND DATEDIFF(month,”r.a_expiration_date”,DATE(NOW)) <= 1 AND r.r_version_label = ‘Expired’
ORDER BY 3

40. Query to find the file system path location of a document
select doc.r_object_id, doc.object_name, MFILE_URL(”,-1,”) as mypath,doc.i_folder_id from dm_document doc where

concurrent_sessions vs max_session_count

concurrent_sessions in server.ini :

The concurrent_sessions key controls the number of connections the server can handle concurrently. This number must take into account not only the number of users who will be using Documentum concurrently, but also what kinds of operations those users will
be executing. Some operations require a separate connection to complete the operation.

For example:
• Issuing an IDfQuery.execute method with the readquery flag set to FALSE causes an internal connection request.
• Executing an apply method or an EXECUTE DQL statement starts another connection.
• When the agent exec executes a job, it generally requires two additional connections.
• Issuing a full-text query requires an additional connection.

The default value is 100. Consider the number of active concurrent users you expect, the operations you think they will be performing, and the number of methods or jobs you execute regularly, and then modify this value accordingly. The maximum number of concurrent sessions is dependent on the operating system of the server host machine. The limit is:

• 1022 for Solaris and AIX platforms
• 2046 for HP-UX
• 1024 for Windows systems

max_session_count in dmcl.ini :

It determines the maximum number of sessions a client can have.

Tracing in Webtop v6.5

Below URL is used to enable tracing on the client side...IP/Localhost(if app server on same m/c) in place of xxx

http://xxx.xxx.xxx.xxx:8080/webtop/wdk/dmclTrace.jsp?level=10&logfile=C:\dfc_trace.log

Also, add the below information in dfc.properties file at the end

dfc.logging.level_to_force_stack=ALL
dfc.logging.verbose=true
dfc.tracing.enable=true
dfc.tracing.include_rpc_count=true
dfc.tracing.include_rpcs=true
dfc.tracing.log.level=ALL
dfc.tracing.print_exception_stack=true
dfc.tracing.verbose=true
dfc.tracing.include_session_id=true
dfc.tracing.mode=standard
dfc.tracing.timing_style=date
dfc.tracing.dir=c:/temp/trace

____________________________________