Renditions in Documentum


A couple of days ago someone had asked me about Renditions in Documentum, Let me try to explain very briefly what renditions are.

Renditions are Representations of a Document in a different format.  In other ways, the only difference between a document and its rendition is only its format.   When you create a document in Documentum you specify the file format for that object, this file format is the Original format of that object.  The renditions of this object have the same properties of the original format object but it’s in a different format. These renditions are made either using Documentum Media Transformation Service or the supported Converters.    

The Converters are basically for Transform one type of graphic image to another type (e.g.: TIFF to JPG) or one-word processing format to another word processing format (e.g.: PDF to MSWord). There are few converters that come with the Content server, you can either buy or make your own custom converters. I am not getting into details of Converters here.

When you request for a rendition that uses converters Content Server creates and manages the Renditions automatically. You can also add renditions manually.
If you wanted to add renditions manually you can create the rendition by your own means and use addRendition() IDfSysObject interface.
For removing a rendition use removeRenditon() of the same IdfSysObject interface.
You have to have at least write permission to remove rendition of a Document.   

DFS Data Model Explained


Hi All 

As I am exploring the new DFS  I am noting some points that I can share with you guys. I know I cannot explain everything about DFS here As like you I am also learing but I will try to give an overall  picture of What DFS is all about here. Please visit this blog often as I keep updating this.

 

DFS Explained

Documentum Foundation Services the SOA face of Documentum. The remote invocation of DFS is implemented using SOAP based web-services. The beauty of DFS is that though DFS is exposed for the remote usages as Web Services it can be called locally using DFS Client libraries. DFS allows the migration easy by allowing services to be developed on existing BOF (Documentum Business Object Framework) also.Another interesting point about this is DFS consolidates the numerous inter depended methods into a Single Service. The DFS data model is expressed in both Java client library class and also in Service XML Schemas. This provides a consistent approach to model data that will be exchanged between the various business processes.  As mentioned above the DFS clients can be of two types.

  • Applications (Consumers) written any language that can use WSDL to interact with DFS.
  • Clients written on java that uses DFS Java Client library

 Data Model of DFS  The data passed to and from the Services are encapsulated into DFS Object Model  These are the few important object types on DFS

  • DataPackage
  • DataObject
  • ObjectIdentity
  • Property
  • Content
  • Permissions
  • Relationship

 Lets see what all these does now 

DataPackage 

According to the DFS Reference Guide from EMC The DataPackage class defines the fundamental unit of information that contains data passed to and returned by services operating in the DFS framework.  That means DataPackage is a collection of DataObject Instances, which is passed back and forth by Object Service Operations. In other words when you call services like Create, Get, Update the Data Object Instances are passed using DataPackage class.

It’s like an Envelope for putting various DataObject Instances. Object service operations process all the DataObject instances in the DataPackage sequentially.

 

DataObject

A DataObject is the DFS representation of a Persistent Object in Documentum Repository. DataObject potentially has all the information related to an Object in repository. This includes the Content, properties (Both Single and Repeated), relation with other Objects in repository, Its permissions etc. Due the nature of these objects, Data Objects are very complex.
Optionally these object instances may have instructions to services about how the parts of Objects have to be processed.
The type field of DataObject class represents the underlying typed object type name that corresponds to the Object Instance. For example dm_user, dm_document etc. default type is dm_document. I.e. if no type is specified then service implicitly assign that to type dm_document   

 

DataObject

ProperySet
Content
ObjectIdentity
Permission
Relationship

 The above Figure represents a DataObject.

ObjectIdentity

 As the name says this represents a Unique Object in the Repository. An instance of this class must have the repository name and a unique identifier for that object.  The value that represents a repository object may be any of the following  

  • OBJECT_ID – Contains r_object_id of the Object (Represents both Current Object and Non Current Object)
  • OBJECT_PATH – Contains String expression of the path of the Object in repository
    [Cabinet]/[Folder]/[File_Name]
    (This Represents only Current Object. Since you can create multiple objects with same name in same directory this does not guarantee uniqueness of the object)
  • QUALIFICATION – DQL Snippet that qualify an Object uniquely
    (The DQL Snippet is the Full DQL that follows the keyword FROM. E.G. dm_document where r_object_id=’09xxxxx’)

 Note: During the Creation of the object all you need to populate is only repository name.  

Property

Each Property represents an Object Attribute (Property) in the Documentum Repository. Property Set is a collection, which works as a container of multiple Property Objects

As there are 2 types of Properties for an Object a property can either be a Single property (Single Value Attribute) or Array Property (Repeated Attributes).  Property class has been sub classed to accommodate various data types and they are as follows.  

StringProperty

 
NumberProperty  
BooleanProperty  
DateProperty  
ObjectIdProperty  
ArrayProperty (Abstract Class) ®
  StringArrayProperty
  NumberArrayProperty
  BooleanArrayProperty
  DateArrayProperty
  ObjectIdArrayProperty
Transient Property

These properties are not the part of persistent properties of a repository object. Transient Property can send custom data fields to a service to be used for a purpose other than setting attributes on repository objects.

 

Array Property and Value Action

The Repeating Attributes of an Object type is represented as Array Properties. As you can see in the above table The ArrayProperty is an array of the corresponding single property.  Another Interesting aspect of Array Property is Value Action. Now lets see what value action is all about? This is an optional Action – Index mapping pair. These pair has the instruction as to what is the action to be done on a particular element of the ArrayProperty. The Index is the position in the attribute and Action is what is the action to be performed on that element of that index.  The possible ActionTypeValues are  

  • Append
    When processing ValueAction[p], the value at ArrayProperty[p] is appended to the end of repeating properties list of the persistent repository object. The index of the ValueAction item is ignored.
  • Insert
    When processing ValueAction[p], the value at ArrayProperty[p] is inserted into the repeating attribute list before position index. Note that 1, which must be accounted for in subsequent processing, offsets all items in the list to the right of the insertion point.
  • Delete
    The item at position index of the repeating attribute is deleted. When processing ValueAction[p] the value at ArrayProperty[p] must be set to a empty value.
  • Set
    When processing ValueAction[p], the value at ArrayProperty[p] replaces the value in the repeating attribute list at position index. (From DFS Development Guide)

 

Content

 Content Class or its Subclass instance represents the actual File Content associated with a DataObject.  DataObject can have zero or more Content objects. It can hold Renditions also.   Content class object can be configured to hold following

  • The Whole Document

  • Page of a Document

  • Pages (One or More) which has been represented by the characteristic

 

 

Permissions

 A DataObject has list of Permission Objects, which decides the permission of the object that’s represented by the DataObject. The permission list provides read access to the permission on an object in repository by the user who has been logged in currently. 

Also have to note an interesting point here that You Cannot change permissions on a repository object by changing the Permission list and saving the DataObject.

For changing permissions of a repository object the real ACL of that object has to be edited or replaced.   As mentioned above multiple Permission Objects creates Permission List. The Permission Object has a field named permission Type that sets what type of permission is set. The values possible for these fields are BASIC, EXTENDED or CUSTOM   Another important thing about permissions are if you assign one particular permission for one user to one object, all the lower lever permissions below the assigned permissions will be granted to that user. This is known as Compound or Hierarchical Permissions  PermissionType Enum Constants (From DFS Guide) 

PermissionType

Permission Description
Basic NONE No access is permitted
  BROWSE The user can view attribute values of content
  READ The user can read content but not update.
  RELATE The user can attach an annotation to object.
  VERSION The user can version the object.
  WRITE The user can write and update the object.
  DELETE The user can delete the object.
Extended X_CHANGE_LOCATION The user can change move an object from one folder to another. All users having at least Browse permission on an object are granted Change Location permission by default for that object.
  X_CHANGE_OWNER The user can change the owner of the object
  X_CHANGE_PERMIT The user can change the basic permissions on the object.
  X_CHANGE_STATE The user can change the document lifecycle state of the object.
  X_DELETE_OBJECT The user can delete the object. The delete object extended permission is not equivalent to the base Delete permission. Delete Object extended permission does not grant Browse, Read, Relate, Version, or Write permission
  X_EXECUTE_PROC The user can run the external procedure associated with the object. All users having at least Browse permission on an object are granted Execute Procedure permission by default for that object.

 

 

Relations

Relationship allows creating a Single DataObject that specifies all the relations that it has with other objects existing and new. This also allows single service call to get, update or create the whole set of objects and its relationships.  Relationship has two subclasses             
Object Relationship  
           &           
Reference Relationship
These classes define all the relationships that a (any) object in repository can have.  This represents existing relations, also can add new relations, which can be updated in the repository using service calls.  

According to DFS Developer guide The repository defines object relationships using different constructs, including generic relationship types represented by hardcoded strings (folder and virtual_document); dm_relation objects, which contain references to dm_relation_type objects; and dmc_relationship_def objects, a representation provides more sophistication in Documentum 6. The DFS Relationship object provides an abstraction for dealing with various metadata representations in a uniform manner 

 

ObjectRelationship

ObjectRelationship represents relationship to a new or existing repository object. It can be used by the update operation to either update or create Content objects. This when called with update operation modify/ create the target objects 

ReferenceRelationship

A ReferenceRelationship represents a relationship to an existing repository object and is specified using an ObjectIdentity. This object can (only) be used to create a relationship between two objects. This will not create or update the target objects.

 

 

Download this study note (PDF)

Changing the Repository Owner Password in Documentum


Follow thse steps to change the Repository owner password in Documentum Content Server

  1. Stop the Repository
  2. Go to $DM_HOME/dba/config/<repository name>/
  3. Create a copy of the file with name dbpasswd.txt
  4. In the DB change the password of Repository Owner
  5. Edit the dbpasswd.txt( The file backed up in Step 3) and replace the new password as plain text
  6. Save dbpasswd.txt
  7. Go to $DM_HOME/bin run the following
  8. dm_encrypt_password -docbase <docbase name> -rdbms -encrypt <database password>
  9. Restart the Repository

What is the significance of i_vstamp and i_is_replica


These are the properties defined in Persistent Object Type and inherited to all its subtypes

i_vstamp
This property is basically used for versioning, each time you save changes to the object the value of this property increases by 1 and this also helps to check the concurrent modification of object.

i_is_replica
this property of object that indicates whether that object is replica of  an object in a remote repository

Object Attribute :- Repeated and Single Value Attributes in Database


Documentum Object support 2 types of Attributes

They are 

  • Single Value Attributes
    Single Value properties are properties that can hold a single value at a time
    For  Eg: object_name   One object will have only a single name
  • Repeated Attributes
    Repeated Attributes are properties that can hold multiple values
    For Eg: Keywords  A object can have multiple keywords attached to it

Documentum Uses Underlying database to save the meta-data. ie the Attribute information, Each attribute type in Documentum essentially has 2 tables in the database. One for saving the single value attribute and another one to save repeated attribute.  Documentum uses the following naming convension for the tables,  object_name_s for single value attribute table and object_name_rfor repeated attribute table where object_name is the the name of object type. for example for the dm_sysobject the tables names would me dm_sysobject_s and dm_sysobject_r

Repeated Property

Each property has a column in the _r table with I_position where I_position determines which location it belongs to 
eg:   if my_object has 2 repeated attributes given_names and location the my_object_r will have following columns r_object_id, i_postion, given_names, location
 

Single Value Attribute
Single Value Attribute is defined as each attribute has single column in the _s table
eg: if my_object has two single attributes surname and sex r_object_id, surname, sex  columns in the my_object_s will be r_object_id, surname, sex

View :- Each object type has 2 views
_SV & SP – All Single value attributes in conjunction with all super types
_RV & RP – All Repeated value attributes in conjunction with all super types
eg: if my_object is extending dm_document views will have all the single / repeated attributes of dm_sysobject,dm_document and my_object (dm_document will be added only if any new attributes has been added to it ) by default dm_document does not have any single attribute or repeated attribute


What is the Difference Between Cabinets and Folders


When I started with Documentum one among the first confusions that I had was this.  Let me explain how are they different.

Cabinets
A Cabinet object is the highest level of Organization visible to the users in a Repository. Means Cabinets can exist only in the top level in the Repository. Cabinet object extends from Folder. But Any object that extends from dm_sysobject (ie means any object that can be saved in a repository) are stored in Cabinets. A Cabinet cannot be placed inside another cabinet or a folder and you need super user privileges to create or remove a Cabinet. But a user who has write access to cabinet can modify the attributes of a cabinet.
The Internal name of Cabinet is dm_cabinet and the r_object_id of cabinet always starts with “0c”

For Understanding Cabinets better,  you can compare the Drive letters of your Home PC as a Cabinet. A Drive is always in the top level in the hierarchy. and you cannot save any data above the Drive level but you can create files in a Drive or in any folders located in the drive.

Folders
The Folders are means of organizing objects that can be saved (Any object that extends from dm_sysobject other than dm_cabinet) and it is used with combination of a Cabinet. In other words a Folder cannot exist without a Cabinet. All objects that are saved in a repository must be saved either in a folder, which is inside a cabinet, or it can be saved directly in a cabinet.
Folder extends from dm_sysobject and the internal name of folder is dm_folder and r_object_id of a dm_folder starts with “0b”

An object has to be linked with a folder and it can be linked with multiple folders