Database Reference
In-Depth Information
objects), and
workflow tasks
(which communicate with other processes to
run packages, send and receive messages, send e-mail messages, and so on).
Creating a control flow in Integration Services requires the following
steps:
Adding containers
that implement repeating workflows in a package.
Adding tasks
of the kinds mentioned above. If the package has to work
with data (which is most of the times the case), the control flow must
include at least one data flow task.
Connecting containers and tasks
using precedence constraints. Tasks or
containers can be joined in a control flow dragging their connectors
from one item to another. A connector between two items represents a
precedence constraint, which specifies that the first one must be executed
successfully before the next one in the control flow can run.
Adding connection managers
, which are needed when a task requires a
connection to a data source.
A
data flow
extracts data into memory, transforms them, and writes them
to a destination. Integration Services provides three different types of data
flow components as follows:
Sources
, which extract data from data stores like tables and views in
relational databases, files, and Analysis Services databases. Integration
Services can connect with OLE DB data sources, like SQL Server, Oracle,
or DB2. Also, sources can be Excel files, flat files, and XML files, among
other ones.
Transformations
, which modify, summarize, and clean data. These trans-
formations can split, divert, or merge the flow. Examples of transforma-
tions are
Conditional Split
,
Copy Column
,and
Aggregate
.
Destinations
, which load data into data stores or create in-memory data
sets.
Creating a data flow includes the following steps:
Adding one or more sources
to extract data from files and databases and
adding connection managers to connect to these sources.
Adding the transformations
to satisfy the package requirements.
Connecting data flow components
.
Adding one or more destinations
to load data into data stores.
Configuring error outputs
on components.
Including annotations
to document the data flow.
We illustrate all these concepts in Sect.
8.5
when we study the implemen-
tation of the Northwind ETL process in Integration Services.
Search WWH ::
Custom Search