Extended data sources are modeled objects
in the metadata repository that represent different data sources.
By creating and importing extended data sources, you can capture
metadata that is not written to disk or that for some other reason
cannot be imported into the catalog. You can then use the extended
data sources in extension mapping documents in order to track and
report on the flow of information to and from the extended data sources
and other assets.
You define extended data sources in a comma-separated value (CSV)
file. During the import process, the metadata of the extended data
source assets is imported into the catalog. However, the CSV file
itself is not imported into the catalog.
Types of extended data sources
There are
three major types of extended data source assets:
- Applications
- Stored procedure definitions
- Files
You have flexibility to use each type of extended data
source to represent the various types of metadata in your enterprise.
The following definitions include suggested usage.
- Application
- Represents a program that performs a specific function directly
for the user or, in some cases, for another application.
- For example, an application might be a database program, communication
program, or SAP program that interacts with a corporate database.
You can use an application extended data source to loosely model SAP
or other data programs.
Applications can have one or more object types.
- Object type
- A grouping of methods or a defined data format that characterizes
the input and output structures within a single application. For example
an object type could represent a common feature or business process
within an application.
- An object type belongs to a single application. An object type
can have multiple methods. The identity of an object type is application.objectType.
- Method
- A function or procedure that is defined within an object type
to perform an operation. Operations pass or receive information as
input parameters or output values. For example, a method can be a
specific operation or procedure call for reading or writing data through
the application and object type. You could also use a method to represent
the equivalent of a database table, while the input parameters and
output values represent columns of the table.
- A method belongs to a single object type. A method can have multiple
input parameters and output values. The identity of a method is application.objectType.method.
- Input parameter
- Input parameters are the most common way to deliver information
from a client to a method. Methods require information from the client
to perform their intended function. This information can be in the
form of presentation options for a report, selection criteria for
data to be analyzed, individual columns, or many other possibilities.
For example, MONTH and YEAR could
be input parameters for a method that analyzes monthly sales data.
- An input parameter belongs to a single method. The identity of
an input parameter is application.objectType.method.inputParameter.
- Output value
- Methods retrieve and return data to the client or application
in the form of output values. Output values can represent the returned
value for the database column or data file field. For example, JANUARY and 2000 could
be output values for a method that analyzes monthly sales data.
- An output value belongs to a single method. The identity of an
output value is application.objectType.method.outputValue.
- Stored
procedure definition
- Stored procedures are routines that are available to applications
that access database systems, and are stored within the database system.
Stored procedures consolidate and centralize complex logic and SQL
statements, and might update, append, or retrieve data. For example,
stored procedures are used to control transactions as condition handlers
or programs, and in some cases are similar to ETL transactions when
they update.
- The extended data source that represents stored procedures is
called a stored procedure definition to distinguish it
from the stored procedure assets that are saved by IBM® InfoSphere® DataStage® and QualityStage®.
A stored procedure definition can have multiple in parameters,
out parameters, inOut parameters, and result columns.
- In parameter
- An in parameter carries information that is required for the stored
procedures to perform its intended function. For example, variables
passed to the stored procedure are in parameters.
- An in parameter belongs to a single stored procedure definition.
The identity of an in parameter is storedProcedureDefinition.inParameter.
- Out parameter
- An out parameter represents the value or variable returned when
a stored procedure executes. For example, a field included in the
result set of the stored procedure can be an out parameter.
- An out parameter belongs to a single stored procedure definition.
The identity of an out parameter is storedProcedureDefinition.outParameter.
- InOut parameter
- You use inOut parameters when a stored procedure requires information
from the client to perform its intended function, and then processes
and returns the same information. For example, an inOut parameter
could be a variable that the stored procedure processes or aggregates
and returns to the calling application.
- An inOut parameter belongs to a single stored procedure definition.
The identity of an inOut parameter is storedProcedureDefinition.inOutParameter.
- Result column
- Result columns represent the returned data values of a stored
procedure, when it queries data or processes data in a database.
- A result column belongs to a single stored procedure definition.
The identity of a result column is storedProcedureDefinition.resultColumn.
- File
- A file represents a storage area for capturing, transferring or
otherwise reading data. Files are often the source of ETL transactions
and can be loaded and moved by using FTP. The extended data source
type file represents files that cannot be imported into
the catalog by standard means. Files that can be imported into the
catalog are called data files, and are not extended data
sources.