What's New in Release 3.0.0

With the release of Powerdesigner Modeller 3.0.0.rc2 we've made a variety of functional and usability improvements.

Summary of changes

Feature	Summary
Model Templates	Ability to create a new model with all predefined settings applied.
Physical Delete	Physical delete support in Unified Anchor Models.
Export	Export specific models.
Code Generation	Changes in code generation and installers.
GDAL Generation	Generation of Generic Data Access models.
Other Changes	Other minor improvements and changes.

Model templates

Creating a new model for each of the i-refactory specific layers is now supported with model templates. Model templates will bring the advantage of using the same settings which leads to higher level of standardisation, reuse and comprehension. The tedious work of manually setting values for default and standard properties is now taken care of with a model template.

Activate

Powerdesigner menu Tools → General options → Model Creation:

Set the path to a shared drive where modellers can use the template models.

Use

From the Powerdesigner menu File → New Model → Templates you can create a new model based on each of these templates:

Each of the templates has predefined settings for:

Model options
Database generation
A default schema and owner
Model specific i-refactory settings such as the proper architectureLayer, staging template, ...

We have shipped 5 template models which you should customize to your specific needs. If you need assistance in customizing the templates to your specific needs please contact I-Refact.

Physical delete

In a central facts model we've added 2 additional properties on each anchor which will be added automatically to your existing model when you open it and the properties do not yet exist.

These additional properties register the state of an anchor regarding it's physical status. If an anchor is considered physically removed the property acm_physically_removed_ind is set to true and the property acm_physically_removed_by contains the modifier responsible for the physical delete.

The properties are set to false, null respectively if the anchor exists (or is recreated).

{info} During the upgrade to release 3.0.0 the anchor tables in the database are automatically extended with these new columns. The metadata for the anchor tables is automatically updated as well.

Impact for a generic data access model

The property acm_record_ind in the GDAL model from now on will be related to the domain: acm_crud_record_indicator. This domain allows for setting the acm_record_ind to a value 'R' or 'D' for respectively setting the record to logically removed or physically removed.

This new domain is automatically added to the generic data access interface model as soon as you open it.

However, you should manually change the domain for the existing acm_record_ind columns. This can easily be changed in Powerdesigner with the "List of attributes" action, apply a filter on the acm_record_ind column, select all rows and choose to change the domain to acm_crud_record_indicator.

{info} When upgrading to release 3.0.0 the domain setting for the attribute acm_record_ind for each generic data access model will be updated automatically. The views, stored procedures, ... will be automatically regenerated as well.

Export

We have changed the behaviour of exporting DDL and metadata for the models in your opened workspace. You now have the option to set an export path, choose with models you would like to export and for each model the option to export DDL, metadata and/or schema's.

Selecting models and options

From an open model open the context menu and choose the menu item "Generate DDL and Metadata":

This results in a window that shows all your Workspace Opened Models:

{warning} You should open all the models in your workspace. Models may have dependencies with other models such as mappings and relationships. If you export a model and an associated model in your workspace is not opened Powerdesigner is not able to resolve the shortcuts to the appropriate objects resulting in erroneous generated DDL and metadata.

You have the ability to:

Change the directory to which the code should be generated.
Choose to auto generate a new directory which will be created in the directory as specified in the IREFACTORY_EXPORT_PATH named variable.
Choose not to generated DDL and metadata for a model by removing it from the list. To refresh the list of all opened models click the refresh button.
Decide which items should be generated: DDL, Metadata, Schema's.

For each of the selected models the requested code will be generated.

Code generation

We now have the capability to generate code and install scripts without the restriction that only a predefined set of databases are supported (the typical default database for each layer). The code generators now take into account the database that is attached to a model.
An install.sql script is generated in the root export directory. The contents of the scripts will execute the install for each of the exported models.

The script execution is wrapped in a transaction. If one of the install steps fails the transaction is rolled back. So no intermediate and inconsistent state is left behind.

{info} The install.sql script is overwritten each time you run an export but only if at least one model is chosen for which to export the DDL script. The install.sql script will only execute the install for the models for which you have chosen to export DDL code.

The install script will not drop existing objects, always creates objects, despite the fact if they already exist, and doesn't take care of delta deployments.

You should remove existing objects or install to a clean environment. Or use a tool like Redgate to create a delta deployment script, which is our recommended tool of choice as it is easy, fast and reliable in creating delta scripts.
A powershell install script is generated as well. It's a simple though effective way to install via Powershell. It will allow for installing on Linux systems. For example: install the DDL on a Microsoft SQL Server Linux Docker container.
All filenames are from now on generated in lowercase. This enables installing on Linux systems with expected behaviour.
The installer scripts from now on use a forward slash to allow for installs on Linux systems.
For GDAL models DDL files were generated despite the fact that these files did not create any objects. These files are not generated anymore.

GDAL Generation

When generating a GDAL model from a CFPL model we will use the GDAL template so we will get a consistent markup of GDAL models.
A physical diagram and symbols for each of the generated entities and relationships is now generated.
If a CFPL model contains a generalisation anchor the generalization anchor is now generated.
A GDAL object is now generated for a Generalisation Anchor.
The domain acm_crud_record_indicator is added to a GDAL model and will be used as the domain for the attribute: acm_record_ind.
A check is added on auxiliary columns: they should participate in a relation.
Creating references takes into account the owner of the parent entity. But for links the reference names in the CFPL still should have a unique name.

Other Improvements and Changes

Indexes on technical staging models

You now have the ability to create indexes on technical staging models. Well, at least the are created in the DDL generation.

{warning} Unique / primary key indexes.

You should not create unique indexes as this will prevent loading data which contains duplicates.

We also do not automatically create a unique index if you create a primary key on a table in a technical staging model.

Primary keys in technical staging models

You now have the ability to create a primary key on technical staging models. The primary key is not physically implemented but is exported as metadata.

The physical not null constraint which is automatically added by Powerdesigner on primary key columns will not be created either (as this prevents loading data which violates the primary key constraint).

Checks

Check	Description
Cyclic mappings	A new check is added which checks if cyclic mappings exists between helpers. An example is when you create a helperA which is a source input for helperB and add helperB as a source for helperA.
Not allowed to map to surrogate ID	When you map to the identity column of a anchor you will now get a check error when executing the model checks.
Auxiliary columns should be a reference	When you validate the GDAL model you will get a check error if an auxiliary column on a table exists which does not participate in a relationship join. Auxiliary columns in a GDAL should only exists as a result of a relationship to a parent entity and therefore should always participate in a join.
Identity setting	A new check is added for surrogate key columns. The Identity setting should be true for all anchors not being a subtype and false for all other entities. The check will return an error if this constraint violates.
Parent Key of Reference	Each reference should have a parent key set in its properties. This constraint is checked during the model validation and will result in an error in case the parentKey of a reference is not set.
Datasource read only	For mapping between Technical Staging and Logical Validation and between Logical Validation and Central Facts models the datasource setting is now checked to be read only.
One table source	A context entity in a central facts model can be mapped to by only one LVL or CFPL Helper source entity. If a context entity is mapped to by more than one LVL and/or CFPL Helper entity an error is reported during model validation.
Column should have a mapping	Isn't executed anymore for helper entities.
Satellite should have a column	Now takes into account reference columns as well.

Changes

Changes in metadata properties for a model

We no longer export the attributes: "DDL Version", "RDBMS Version" and "Metadata Export Version" on model (interface) level.

Instead we now export the following properties: "Exported for irefactory schema version", "Requires irefactory application version", "Created with irefactory modeller version".

With each release of the Powerdesigner UAM Modelling components we will internally set the proper values for these metadata properties.

Icons in a Central Facts model

The icons for temporal/bitemporal context have changed.

An immutable context has no icon for temporality, a transactional context has a black clock, a bitemporal context a white clock.

References for derived anchors or context

References in a CFPL are now created for derived anchors and context as well.

Upgrade

Upgrade Notes

Constraint violation actions are applicable to certain constraint categories. Not all combinations of constraint categories and violation actions are allowed.

An attribute must have a value, whatever that value may be. It must not be NULL.

A data type of an attribute defines what value an attribute can hold. The data type specifies what type of mathematical, relational, or logical operations can be applied to it without causing an error.

An attribute datatype constraint is the most basic constraint type. It checks for the datatypes we support and have implemented.

For example, we check for string, length of string, integer, date, etc. In the following figure you can see the supported data types by PowerDesigner.

Image is omitted: Supported data types

Constraints can be violated and there are some actions that can be performed when a violation occurs. The possible actions are: EMPTY COLUMN, NO ACTION and SKIP ROW.

An attribute value constraint is an expression that is evaluated. The person who writes the expression is responsible for the correctness of it. The expression should be formulated in a positive way and lead to a Boolean answer. If the expression validates to True, than the value is correct.

Examples

The values in attribute X has to be bigger than 10: X > 10
The email address has to be in a certain pattern: email address LIKE '%_@_%.__%'

A Concept Integration Model is also a central facts model on which you place integration patterns. It is not required to create a concept integration model, but it can be very useful.

Every constraint is assigned to a constraint classification.

The main purposes of the Generic Data Access Layer (GDAL) are to provide logical perspectives for data consumption and to manage CRUD actions.

A generic data access model is a virtual data model that acts as an interface bridge between consumer applications and the central fact storage.

Every attribute is assigned to an attribute classification.

An entity record constraint checks whether an attribute meets the requirements set by another attribute belonging to the same entity.

The main purpose of the Logical Validation Layer (LVL) is to transform the data received from external data sources to fit into the logical data model structure. It is also responsible for validating deliveries. The Logical Validation Layer is also known as the Historical Staging In (HSTGIN) Layer.

The logical validation model is the representation of a single external data source in a logical format. It represent how data delivered by a specific tenant should be transformed, temporalized and validated in the {popup}logical validation layer. The logical validation model is also known as Historical Staging model (HSTGIN).

Multi-active attributes are attributes that contain a business key to provide multiple context records at the same time. For example: a customer has multiple types of phone numbers. “Home”, “Work” and “Mobile”. In that case we add a dependent entity on customer with key “Phone Nbr Type”. This is to prepare for the CFPL multi-active key on customer.

The main purpose of the Technical Staging Layer (TSL) is to create a common starting point for further data processing. It receives data delivered from external data sources and temporally stores them in a database. The input data should be in a tabular format (rows and columns).

Bi-temporal attribute is an attribute that changes over time: they follow a valid timeline. For example, a Part may have a price valid for December and a price valid for January.

Every entity is assigned to an entity classification and to a parent entity classification. The possible values for entity classification are: ALTERNATE KEY CONTEXT, ATTRIBUTE CONTEXT, GENERALIZATION,HELPER, REFERENCE CONTEXT, STABLE, STABLE DEPENDENT and STABLE INDEPENDENT

Entity Set Constraint An entity set constraint can be used to perform a check concerning values of two or more attributes that belong to different entities or to perform a check concerning the value of an attribute with respect to a set of values.

A Set Constraint Helper is a helper in the logical validation model. It is the implementation of a set constraint. The helper returns the records of an entity for a given set constraint, where the instances of this entity do not meet the definition of this set constraint.

The business requirements describe how data should be delivered for the data consumers (end users or applications) in terms of concepts, relationships between concepts and constraints to validate the data. These requirements can be described in a logical data model, for example.

A Business Rule Helper is a helper in the central facts model. It is a set-based calculation of derived facts. You need to use a Business Rule Helper if you want to make a calculation and want to keep a transaction history of the results of this calculation. You use the existing entities from the central facts model as input. The results of the helper must be materialized in 'regular' fact entities, such as Anchors and Contexts, to make them accessible in the Generic Data Access Layer.

Closed Open means that the timeline is valid from (vanaf in Dutch) the supplied valid start date until - but not including - (tot in Dutch) the supplied valid end date. In practice, this means that the start date of a valid time record is equal to the end date of the previous valid time record.

You need to create context-based entities when a set of data may be delivered within the boundaries of a parent context. A context-based entity applies when:

At least 2 entities are delivered.
A context relationship exists between these 2 entities. One entity is the parent context of the other entity.
The parent context entity is delivered as a delta and the child entity is delivered as a full set.

You need to create context-based entities when a set of data may be delivered within the boundaries of a parent context. A context-based entity applies when:

At least 2 entities are delivered.
A context relationship exists between these 2 entities. One entity is the parent context of the other entity.
The parent context entity is delivered as a delta and the child entity is delivered as a full set.

The Management Model contains the PowerDesigner objects for the Unified Anchor Modelling (UAM). When a UAM object is created, a so-called PowerDesigner replica of the corresponding Management Model object is created. This means that certain properties such as metadata columns and column stereotypes are configured in the Management Model and cannot be changed. The replication settings specify which elements of an object can be changed after creating a replica from the template object. It is possible to override the replication settings of an UAM object and change a specific property.

The temporal atomic type describes the datatype of the temporal attributes|

The main purposes of the Central Facts Layer (CFL) is to store data historically. It can also integrate data from different sources. The Central Facts Layer is also known as Central Facts Persistency Layer (CFPL)

The central facts persistence implementation model is the representation of facts in an anchorized data model with the ability to integrate multiple logical models.

In the context of i-refactory, data transformation refers to operations involved in turning raw data readily useful and closer to the business requirements.

Integration patterns are used to integrate entities from different data models. If two or more entities from different data models share the same business key, you can use the Integration Pattern named Key Root. It is a good practice to capture integration patterns in a separate model, named Concept Integration Model.

An attribute is mandatory when its value can not be empty (NULL).

A Physical Data Model (PDM) represents how data will be implemented in a specific database.

{note} The i-refactory uses four PDMs: technical staging model, logical validation model, central facts model and generic access model. Each one of these models is implemented as an additional database, which is used to store data from external and internal data sources.

Reverse engineering is the process of reconstructing a physical and/or Entity Relationship (ER) model from an existing data source. The purpose of reverse engineering is to avoid manual work as much as possible.

Architecture layer

The core of the i-refactory architecture has four layers: TSTGIN, LVL, CFL and GDAL. There are also two auxiliary layers: UCLVL and EXT.

If an entity has one or more attributes that changes over time and you want to keep track of when a attribute is valid at a certain transaction time, then you have a special case of a regular dependent entity, called bi-temporal entity. The bi-temporal entity stores historical data with two timelines. The primary key of the bi-temporal entity is composed by the primary key of the parent entity and the valid start date attribute. The attribute that changes over the valid time is called a bi-temporal attribute.

A delivery agreement is a contract between a Tenant and a Logical Implementation Model or Generic Data Access model. An agreement has a duration. The delivery agreement set the architecture layer (interface) where the data should be ingested as well as the default settings to be applied to the deliveries.

A dependency mapping is a mapping between a helper (or BR helper) and a source entity used in the query of the helper. The helper and the source entity must belong to the same model.

Default dependency is set on entity level (source entity to helper entity)
To allow lineage on attribute level, via the Mapping editor, you could manually add the dependency on attribute level.

An Independent Entity is an entity that implements an Anchor for a Business Key that ‘stands alone’ e.g. that does not contain a reference to another Entity.

A Logical Data Model (LDM) matches the language, structure and quality of the business, regardless of the physical data implementation. The Logical Data Model reflects the business requirements.

A delivery may be considered as "untrusted" if deletes of data in the Logical Validation Layer have taken place and the processing of new deliveries cannot 100% rely (trust) on having enough statistics and data available to detect logical deletes, to determine the exact delta and to execute set based validations.

A Dependent Entity is an entity that implements an Anchor for a Business Key that ‘depends’ in its existence on another Entity. A Dependent Entity contains Business Key fields of which at least one is a foreign key (FK).

The transaction time in i-refactory is different from what is commonly understood by transaction time. Transaction time is usually seen as the moment when a fact was stored in the database. In the i-refactory, the transaction time is the time, as dictated by the source system, not by the i-refactory database.

The Attribute type links the attribute to one of the existing interfaces.

Computed columns are columns whose content is computed from values in other columns in the table.

Functional date A functional date or time is a point in time and is defined by a user. An example is an order date or date of birth.

The technical model (also known as Technical Staging In model: TSTGIN) is a representation of how exactly one delivery from a specific data source will be processed in the technical staging layer.

Generalization is the process of extracting shared characteristics from two or more classes (hyponyms), and combining them into a generalized superclass (hypernym). For example: an 'employee' and a 'customer' are both 'persons'.

The Mapping Editor provides a graphical interface for creating and viewing mappings between models. It provides a global view of all the mappings related to the entities of a given model, allowing you to quickly identify those which are mapped and not mapped.

When a certain fact can change over time and you need to capture when that fact is valid in the real world, you can add a valid start date and a valid end date to the entity.

A valid time tells us in which period a record is valid. While a functional date represents just one point in time, the valid time has a begin and an end date, for example:

For Order item 123, a Retail price of 10.00 was valid from 2019-01-01 to 2019-06-01.
For Order item 123, a Retail price of 12.00 was valid from 2019-06-01 to 2020-01-01.

Alternate key is an attribute or a group of attributes whose values uniquely identify every record in an entity, but which is not the primary key

Candidate key

A candidate key consists of one or more attributes and meets the following requirements:

Unique: The value of the key defines uniquely one instance of a concepts. There are no double values.
Non-volatile: (Almost) doesn't change.
Minimal: Contains only the elements needed.

There are two kinds of candidate keys:

primary key
alternative key

A Key Root Link is an integration concept that must be used when the exact same business concept or dependent business key occurs in different models. The Links for this dependent business key in the different UAM models are all subtypes of the Keyroot Link.

Normalization is the process of decomposing tables in a database in order to reduce data redundancy and improve data integrity.

A strongly typed model is a model in which each all attributes have a predefined data type, for example: integers, doubles, date.

Surrogate Key A surrogate key is a system generated unique identifier that does not have any contextual or business meaning.

Business Key

A business key is an unique identifier that has business meaning and exists in the real world outside of the database. It consists of a column or a set of columns that already exists in a table. A business key is also known as a natural key

A Key Root Hub is an integration concept that must be used when the exact same business concept or independent business key occurs in different models. The Hubs for this independent business key in the different UAM models are all subtypes of the Keyroot Hub.

A relationship shows how two entities are related to one another. For example, a customer can place an order, and a order can have a customer.

Every Attribute has an atomic type (data type) which is linked to the attribute type of that attribute.

The cardinality shows how many instances of an entity can take place in a relationship.

An enumeration consists of the list of values that a given attribute should adhere to.

{example} An order can have different statuses, such as shipped,packing,created anddone. Other statuses are not allowed.

Foreign Key

A foreign key is an attribute or a set of attributes that refers to the primary key of another entity. The original entity containing the primary key is called the 'parent' entity and the entity containing the foreign key is called the 'child' entity.

A natural key is an unique identifier that has business meaning and exists in the real world outside of the database. It consists of an column or a set of columns that already exists in a table. A natural key is also known as a business key

The primary key is an assigned key that consists of a minimal set of attributes to uniquely specify an instance of a record. The attribute or a combination of attributes should meet the following characteristics:

Unique: The attribute values of the key uniquely identify one instance of a concept. There are no duplicate instances.
Non-volatile: The key does not change.
Mandatory: All values are filled; there are no NULL values.

It is good practice to choose a primary key that also meet the following characteristic:

Safe: Doesn't contain private or sensitive information, such as a social security number.

Constraints are related to the other elements depending of the type of the constraint. Certain constraints are associated to attributes, entities, helper entities, unique keys or relationships between entities.

An attribute may be assigned to one or more entities (ex: acm_exists_ind) and an entity may have several attributes

Each layer may have one or more interfaces. The amount of interfaces depend on how many tenants and delivery agreements have been configured.

Namespace is what in the terminology of SQL Server is called database schema.|

A Delivery is a container that holds the specification of what is actually pushed to the i-refactory platform. This specification consists of a list of entities.

Key Root A Key Root is a central repository for Business Keys. A Key Root ensures that similar records out of different data sources are identified by both the same Business Key as the Surrogated Key.

Context

A Context is a temporal table with a transaction start and end date. The Context tracks all changes of the context attributes related to a business key in the transaction time. This means that every change of an attribute value in a source system leads to a new record in the Context. The old record is end dated with the load date and the new record is start dated with the load date.

Hyponym is a term that denotes a subcategory of a more general class. For example: 'cat' and 'dog' are a hyponyms of 'animal'.

A mapping establishes relationships between concepts of separate data models. It creates a link between entities and attributes from a source model to related entities and attributes in the target model. A source model should precede the target model in the i-refactory architecture.

oasi_bk is an abbreviation for One Attribute Set Interface (OASI) with business keys. A normal view in the generic data access layer (GDAL) consists of the surrogate key, foreign key and attributes. The oasi_bk-view in the GDAL is a view where the business key(s) are also shown.

A subtype is a subgroup of an entity. You can create a subtype if a group of instances share some attributes and relationships that only exist for that group. For example, entity Customer can have a subtype Company and a subtype Person. They share the common attribute customer number, and can have some attributes of their own. Such as birth date for a Person. The entity Customer is called a supertype.

A subtype:

inherits all attributes of the supertype
inherits all relationships of the supertype
usually has one or more own attributes
can have subtypes of its own

Anchor: Independent Entity

An Independent Entity is an entity that implements an Anchor for a Business Key that ‘stands alone’ e.g. that does not contain a reference to another Entity.

Anchor: Dependent Entity

A Dependent Entity is an entity that implements an Anchor for a Business Key that ‘depends’ in its existence on another Entity.

A domain will help you to identify the types of information in your model. It defines the set of values for which a column is valid. A domain can specify a data type, length, precision, mandatoriness, check parameters, and business rules. It can be applied to multiple columns, which makes it easier to standardize data characteristics for columns in different tables.

Each interface may have one or more entities and one entity belongs to only one interface. An entity belongs to an i-refactory data model.

A helper entity creates a derived entity and can be used when you need to transform, filter, or calculate data. The purpose of a helper differs per model:

Technical model: a helper is used to transform data.
Logical validation model: a helper is an implementation of a set constraint (Set Constraint Helper).
Central facts model: a helper is used for a set-based calculation of derived facts (Business Rule Helper).

HSTGIN is the abbreviation of Historical STaging IN. It is an older term to indicate the Logical Validation Model or Logical Validation Layer.

A schema is a set of database objects, such as tables, views, triggers, stored procedures, etc. In some databases a schema is called a namespace. A schema always belongs to one database. However, a database may have one or multiple schema's. A database administrator (DBA) can set different user permissions for each schema.

Each database represents tables internally as <schema_name>.<table_name>, for example tpc_h.customer. A schema helps to distinguish between tables belonging to different data sources. For example, two tables in two schema's can share the same name: tpc_h.customer and complaints.customer.

A Tenant is a delivering party for a dataset or datarecord as agreed in the Delivery Agreement.

TSTGIN is the abbreviation of Technical STaging IN. It is an older term to indicate the Technical Model or Technical Staging Layer.

An index organizes data in a way that improves the speed of data retrieval from a database. To maintain the index data structure, there is a cost of additional writes and storage space.

The acronym CRUD stands for create, read, update, and delete. These are the four basic functions of persistent storage.

OLAP is a acronym for Online Analytical Processing. OLAP is category of software tools which provide analysis of data for business decisions. It uses complex queries to analyze aggregated historical data from OLTP systems.The primary objective is data analysis and not data processing.

OLTP is a acronym for Online transaction processing. OLTP captures, stores, and processes data from transactions in real time. Its primary objective is data processing and not data analysis.

A hub or independent entity is an entity that implements an Anchor for a business key that ‘stands alone’ e.g. that does not contain a reference to another entity. An independent entity contains business key fields, that show up as alternate key (AK), and the primary key (PK) is its surrogate key (ID).

A key is a combination of one or more attributes of an entity that uniquely defines one instance of that entity.