Understanding SAS/Warehouse Administrator

Michael Davis, Bassett Consulting Services, North Haven, Connecticut

ABSTRACT • Raw operational data is filtered into a sub-set to

remove columns and rows that are not required
Some firms have looked at SAS/Warehouse for typical decision support activities. Filtering
Administrator and decided to continue developing raw data down to the essential items can
applications in the traditional way. Why? Reasons improve the speed of subsequent decision
include high investment cost, difficulty of support activities.
incorporating legacy code, and the awkwardness of
using terminal emulators when running • “Header” and “Detail” files need to be joined. In
SAS/Warehouse Administrator on a server located operational data files, common information such
in a “glass room” or other remote location. as a client’s physical address is removed from
the transactional detail files and placed into a
However, if one can overcome these objections and header file to save disk space. To analyze the
let SAS/Warehouse Administrator write the required transactions, it is often necessary to re-join the
SAS code, the advantages of metadata take over. header information with the transactions in the
These advantages include easier maintenance and detail file.
more rapid development of new data warehouse
applications. Another advantage includes quickly • Tables need to be sorted and indexed.
determining the impact of changing columns and Operational data is often sorted by the keys
rows. Last, SAS/Warehouse Administrator necessary to quickly find a customer
automatically publishes HTML documentation and transaction. It would be a lucky accident to find
process diagrams. that this order is also the best sort sequence to
support decision support reporting. Because
This paper will illustrate how a data mart is modeled tables often need to be accessed by multiple
in SAS/Warehouse Administrator, drawing upon a keys, it is often necessary to create multiple
prototype that the author recently created. The indexes to promote efficient information retrieval.
example will highlight how conditional processing
can be accommodated by CASE expressions. A • Tables need to be summarized. Decision
technique to work around the limitations of terminal support activities often require summarized
emulation will also be demonstrated. data, collapsed by the analysis categories.
Rather than going through the effort and
expense of summarizing the detailed information
WHAT IS DATA WAREHOUSING ? each time a report is run, a better strategy is to
pre-summarize the detailed tables once into the
Before one can make the case for selecting summaries that may be needed.
SAS/Warehouse Administrator as the tool to use
when creating a data warehouse or data mart, it is • OLAP. When multiple summaries of the same
important to define what data warehousing means. detail information is required, disk space and
retrieval performance can often be improved by
The author defines data warehousing as the process storing the summarized information in OLAP
of making operational data available to decision (On-Line Analytical Processing) structures, often
support applications, such as SAS. Data known as “cubes”.
warehousing involves extracting, transforming,
joining, sorting, summarizing, and consolidating • Standardize code schemes. As an example, a
operational data. customer’s sex might be denoted in one table as
either “M” or “F”. In another table, the
information may be stored as 0 or 1. To
INTRODUCTION TO DATA WAREHOUSING facilitate enterprise-wide reporting, it is desirable
to transform disparate code schemes to
The skeptical reader might ask, “Why undertake this common ones.
activity?” To address the concerns of such readers,
one might offer the following common computer • Cleanse “dirty” data. Consider the previous
programming [SAS] activities that are performed in example. If the customer (patient) is coded as
the cause of data warehousing: “M” (male) and pregnant, then it appears that an

1
 effort to correct this and other logical the same data sources, performance of the
inconsistencies should be mounted. Edit-check transaction systems can suffer. A better
programs can identify elements that require solution may be to create or update data
cleansing and can perhaps correct some errors warehouse and mart tables from the operational
without manual intervention. systems during off-peak hours.

• Standardize physical file structures. Some data • Save human resources. In organizations where
may be in flat files, some in SAS data sets, and data warehousing is not well organized, multiple
others in third-party data base management persons often duplicate efforts to transform
systems (DBMSs) such as Oracle, Microsoft operational data for reporting. Some of these
SQL Server, and DB2. To facilitate reporting, persons may not have the appropriate skills or
the data should be transformed into tables of tools to perform this task. One of the benefits to
same physical format. If the reporting is to be data warehousing is to save human effort and
done with SAS tools, then the data should be costs in creating and maintaining data
stored in SAS tables, MDDB cubes, or views warehouses and data marts.
created by SAS/ACCESS.

WHY SAS/WAREHOUSE ADMINISTRATOR ?

BENEFITS OF DATA WAREHOUSING
Some readers might exclaim at this point, “Yes, we
At this point, the skeptical reader might exclaim, “I see the value to data warehousing. But why should
can see the value of data warehousing but how do I we try to convince management to license yet
justify the cost and effort to my management?” another SAS product? Can’t we do data
Here are some common benefits that data warehousing with Base SAS?” Of course
warehousing can yield: organizations can create data warehouses and data
marts with Base SAS and other tools.
• Reduce intra-organization discrepancies. When
each department or division undertakes the However, it is this author’s proposition that when
preparation of raw data for reporting, different one considers the total cost and effort required to
assumptions and techniques can yield different create, maintain, schedule, and document data
results. This can lead to more effort being spent warehouses and data marts, licensing
on reconciling the differences than on what the SAS/Warehouse Administrator may be the least
results mean. expensive alternative. Consider the following
benefits that may be gained by using this product:
• Reproducible results. Operational data often
changes. If one runs the same report later, the • “point and click” interface
results may differ from the first run. When the
data source for a report is a data warehouse or • ability to accumulate, maintain, and report on the
mart table that represents a “snapshot” taken at warehouse’s metadata
a specified interval, the report’s users can count
on consistent results. • control processes across multiple platforms

• Document data repository. Much effort is • changes are automatically posted to generated
consumed in answering questions such as SAS code
“Where does that number come from?” and
“What does that code represent?” One common • process flowcharts can be generated
benefit of data warehousing is that the process
and resulting data stores of the data warehouse • HTML document can be generated and posted
and marts are documented for the benefit of the to a web server
potential users.
• Dependent job scheduling and load-sharing can
• Improve performance of operational systems. be accomplished via the LSF JobScheduler
Transaction systems are often designed to give
best performance when a few records are to be • Process libraries and other features are
retrieved. By contrast, decision support available to structure the warehousing process
applications typically read entire tables. When
transaction and decision support systems share

2
WHAT IS METADATA AND WHY IS IT NEEDED ? Another feature of the Data Warehousing
Community section of the SAS web site that is well
Metadata is information that defines sources, data worth exploring is the “Getting Started with
stores, code libraries, and other resources. It is SAS/Warehouse Administrator”. Also supplied with
used to write the actual SAS code. Technical the software, it can be found at:
metadata defines where the data lives and how to
access it. Business metadata defines what the data http://www.sas.com/service/tutorials/v8/warehous/index.html
means and who is responsible for it.
This tutorial is invaluable to those attempting to set
Perhaps the major advantage of using up their first project in SAS/Warehouse
SAS/Warehouse Administrator is that it facilitates Administrator.
the creation and maintenance of metadata.
Consider the following example. It is decided to
change the logic used to transform a column of STARTING SAS/WAREHOUSE ADMINISTRATOR
intermediate information? How do we find all of the
places affected by the change and make sure that The current version of SAS/Warehouse
they use the new logic? Administrator is run within a SAS session as a SAS
desktop application. The SAS desktop is a graphical
Without a tool such as SAS/Warehouse interface to tools and files. It is supplied with Base
Administrator, making changes to an existing data SAS. However, the SAS desktop is more commonly
warehouse or mart can be a nightmare. Metadata used to access features of SAS/EIS and other SAS
gives us a single point of control, even when products.
warehousing occurs across multiple computer
platforms. To start SAS/Warehouse Administrator, one can
select -> Solutions -> Development and
SAS/Warehouse Administrator facilitates changes to Programming -> Warehouse Administrator.
programs that create and maintain data warehouses However, the author finds it more convenient to
and marts because it actually generates the SAS issue the command “DW” from the command bar.
code to be run. It also provides tools to search, This will open up a window similar to the one shown
report, and document the metadata. Finally, in Figure 1:
SAS/Warehouse Administrator can import and
export metadata to other applications. This opens
the possibility of using additional tools to create and
maintain a data warehouse or mart.

DATA WAREHOUSING VOCABULARY

Figure 1
The term “data warehouse” is commonly used to
describe all outputs of data warehousing. However, To open an existing warehouse environment, one
it is the author’s conclusion that many repositories merely double-clicks on the icon representing that
created by data warehousing are more accurately environment. To create a new warehouse
described as “data marts”. Data marts are environment, right-click on white space within the
distinguished from data warehouses in that they are SAS/Warehouse Administrator window.
organized to support a specialized, specific
application and a finite set of reports.
Environment Hierarchy
The acronym “ETL” stands for extract, transform,
and load. ETL processes represent the major One of the confusing aspects of SAS/Warehouse
activity associated with data warehousing and the Administrator that confronts new users is the
use of SAS/Warehouse Administrator. hierarchy of warehouse elements. The hierarchy is
illustrated in Figure 2, shown on the next page. The
The SAS web site lists other data warehousing following limited hierarchy description may help
vocabulary that may be helpful to the uninitiated. It those beginning to use this product.
can be found on the SAS web site, in the Data
Warehousing Community at: Within a typical warehouse environment, there are
usually Data Warehouses and Operational Data
http://www.sas.com/rnd/warehousing/glossary.html Definition Groups. Data Warehouses are further

3
organized into Subjects, which may contain Data The Count Input ODD Group defines all of the input
Groups, Infomarts, and OLAP Groups. Operational sources to our warehouse. In addition to the files
Data Groups Definition Groups include one or more received periodically from three countries (France,
Operational Data Definitions. Figure 2 shows the Italy, and Netherlands), the Call Center History table
hierarchy of groups and data stores when our also appears. This icon refers to the same physical
“demo” environment is opened: table that the Call Center History data table. This is
because the inputs to warehouse processes must
be defined as ODDs and output tables must defined
as data tables.

GLOBAL METADATA

In our demo, there is just a single warehouse.

However, in practice, we may wish to create multiple
warehouses. It would be a nuisance at best if we
had to define global parameters for each
warehouse. So SAS/Warehouse Administrator
allows us to define in one place all of the metadata
that may be shared across multiple warehouses in
the same environment.

To get to the global metadata in SAS/Warehouse

Administrator, one selects -> File -> Setup… A
window similar to the one illustrated in Figure 3
appears:

Figure 2

This demo environment was created to illustrate how

data from telephone calls made to customer service
centers in multiple countries might be periodically
consolidated and summarized. Figure 3

In our demo environment, we see the HASUG demo One can create, examine update, and delete
warehouse environment icon at the top of the different categories of metadata by selecting a radio
hierarchy. Next in the hierarchy, we see station in the Type window, selecting the metadata
HASUG_demo data warehouse icon and the item in the window below it (SAS Libraries as shown
Country Input ODD (Operational Data Definition) in Figure 3), and clicking on the Add, Edit, or
Group icon. Remove buttons.

Under the HASUG_demo data warehouse icon is SAS library metadata is largely self-explanatory.
the Call Center Data Group, the Temporary Files However, one of the author’s tricks is to use
data group, and the MDDBs subject. In Figure 2, SAS/ACCESS Libname engines to define data
the icons for the history tables and input errors sources in DBMSs as SAS libraries rather than
tables are shown under the Call Center Data Group. DBMS Connections. This trick seems to work better
when the DBMS options are embedded in the Path
Under the MDDBs subject icon, there is a single text box instead of the Options text box on the
MDDBS OLAP group. Under this group are the Details tab.
three MDDB OLAP cubes created by the demo
environment.

4
In our demo, only the computer on which OPERATIONAL DATA DEFINITIONS
SAS/Warehouse Administrator is defined as a host
computer. However, in a distributed computing Operational Data Definitions are metadata records
environment, remote hosts can be defined to and that provide the instructions to access data sources.
can be controlled by SAS/Warehouse Administrator. Figure 4 illustrates an Operational Data Definition
Properties Window.
All elements of a data warehouse have an owner
and an administrator as attributes. The contact
information for all individuals who serve in these
roles is defined in a single place. This makes
updating this information much more convenient.

Last, the information about the scheduling servers is

entered as global metadata. SAS/Warehouse
Administrator allows users to define CRON, AT, and
null scheduling servers. The null scheduling server
writes a file that is used by the LSF JobScheduler
and other third-party scheduling servers.

TYPICAL SAS LIBRARIES

Figure 4
The libraries (librefs) that should be entered into the
SAS/Warehouse Administrator will vary with each The General tab allows one to enter a description
project. Some of the libraries defined as global and indicate the table’s owner and administrator.
metadata will be assigned by SAS/Warehouse The Data tab allows one to specify the host, library,
Administrator. Other libraries may be assigned and table name. The Columns tab, illustrated in
externally, either when the SAS session is started, Figure 5, shows the type information about the
or as part of user-written SAS code. table’s variables (columns) that one would see in a
“contents” listing.
The following libraries are typically assigned as part
of a warehouse environment’s metadata:

• DBMS engine librefs

• Detail Data
• Source Code
• Metalib (_DWMD)
• Process Library (_SASWA)
• Warehouses

The _DWMD and _SASWA are required by

SAS/Warehouse Administrator. The requirements of
add-in tools make it a good idea to assign the
_SASWA externally through the autoexec.sas
program. Figure 5

The author often defines the Work libref as part of Two tips can be shared about the Columns tab. It
the metadata so it can be used when defining can take a bit of time to key in the required
temporary tables that should disappear when the information for a new table. If a similar table already
SAS session ends. If MDDBs are to be used exists or can be generated by running some legacy
outside of SAS/Warehouse Administrator, such as SAS code, it is much faster and easier to “import”
with AppDev Studio or WebHound software, the required information from that table. Also, after
then it may be useful to define an MDDB libref in the moving rows up or down with the arrowhead buttons
global metadata and assign it through the at the bottom of the window, right-click on a row and
autoexec.sas program. select Save Order to retain the new order after the
properties window is closed.

5
At this point, it might be a good idea to define any Table… This brings up the selector shown in
additional ODDs required. After all the ODDs have Figure 7. Select the category of output table to be
been defined, the next logical step is usually to added and click on the Show button to display the
define the required output structures. output tables available to be added.

These include data tables, MDDB cubes, and other

output files. These typically include flat files,
comma-separated value (CSV), and various Output
Delivery System (ODS) destinations. Then it is time
to start defining the transformation of input tables
into target outputs from the Process Editor.

USING THE PROCESS EDITOR

The Process Editor is used to manage jobs, job

flows, and process flows. These properties must be
defined in order for SAS/Warehouse Administrator to
generate the source code for the transformation
jobs. Figure 7

There are multiple ways to bring up the Process Next, for each output table, define the input data
Editor. One method is to select Tools -> Process source(s). This is done from a selector similar to the
Editor from the pull-down menus. Figure 6 shows a one shown in Figure 7. Again the same physical
sample Process Editor window. table can be an output table and an input table
within the same process flow.

MAPPING STEPS

Experienced SAS users may ask, “How do I

transform the information contained in an ODD into
an output table or MDDB cube?” “How do I embed
these transformations within the metadata?” This is
done through mapping steps.

As one might anticipate, mapping steps define how

columns and rows from the input tables are mapped
to output tables or MDDB cubes. Mapping steps
can specify one-to-one, one-to-many, or many-to-
one mappings. SAS/Warehouse Administrator uses
the metadata in the mapping steps to generate
PROC SQL code to effect the transformations.
Figure 6 However, add-in-tools can be used to customize the
behavior of the mapping steps.
The Process Editor window consists of two panes.
In Figure 6, the left pane shows the Job Hierarchy An example of the dialog box that sets the mapping
after it has been partly expanded. Under the jobs step is shown in Figure 8 on the following page. The
are the output tables and files produced by the jobs. contents of the General tab are displayed. On this
tab, and through SAS/Warehouse Administrator, one
The right pane shows the part of process flow for the can add annotations by clicking on the Notes button,
job, output table, or file currently selected in the Job which brings up a notepad window. The notes are
Hierarchy pane. The direction of flow is from bottom saved in a catalog source entry.
to top, left to right.
An example of the Source Code tab is shown in
The author recommends adding the process Figure 9 on the following page. This tab allows one
output(s) first. This is done by right-clicking in the to select whether the SAS transformation code is
Process View pane and selecting Add Output

6
generated by SAS/Warehouse Administrator or is Mapping can be either 1 to 1 (1:1) or derived. If any
written by the user. of the column names are shared between the input
and output tables, clicking on the button labeled
“1 to 1 Mappings…” automatically sets mapping
relationships for those columns whose names
match.

While mapping relationships are often 1:1, SAS

veterans will want to know how they can embed
conditional mapping assignments. Those familiar
with SQL (Structured Query Language) will
recognize the solution, which are CASE
expressions.

Figure 8

Figure 10

Case expressions follow the format CASE…

WHEN… ELSE… END. Consider the following
example used to flag missing values.
Figure 9
case when picktwo_fr.= '' then 'X' else '' end
In mapping step illustrated in Figure 9, a two input
table SQL join has been specified. When an add- One derived mapping requirement that tested the
in-tool has been specified, the Source Code Library author’s creativity was how to code a “left” join. Left
selector will show “Process Library – SASWA” and joins are often required when updating a master
the Catalog Entry Name selector will show the name table. They are required because we only want to
of the catalog entry of the add-in tool. replace (update) information in the master table
when a valid transaction has occurred.
The Execution tab specifies the computer on which
the process is to execute. The Output Data tab The trick to accomplishing this feat is the Coalesce
specifies the location of the output table. An function. Consider the following example:
example of the Column Mapping table is shown in
Figure 10. coalesce(fr.city, history.city)

One advanced use of the Output Data tab is to This expression replaces the value in the master
specify which rows are written to each of multiple (history) table for city with the value of city in the
output tables. This is done by first selecting the transaction (country) table only when city is a non-
appropriate target table by clicking on the down missing value in the transaction table (and when the
arrow. Then while the desired target table is WHERE keys match).
displayed, click on the Generation Options button.
Then on the Row Selection tab, specify “Row As noted earlier, one of the advantages of using
Selection Conditions” or “User Defined Statements” SAS/Warehouse Administrator is that many
to direct the rows to be output to the target table. expressions can be built using a “point and click”

7
interface in lieu of typing, illustrated by Figure 11 on the same result could be accomplished by the
the following page. appropriate specifications to the warehouse
metadata. This temptation should be strenuously
avoided!

User-written code becomes a “black box”. If other

parts of the warehouse are changed or updated, the
user-written load step will not automatically reflect
those changes. The goal in using SAS/Warehouse
Administrator is to model the process within the
metadata and let the SAS/Warehouse Administrator
generate the SAS code.

One last tip about load steps is not to forget to

Figure 11 specify the host on which the load step is to run. It
is very easy to forget this requirement.
While almost everyone appreciates the reduction in
typing offered by “point and click” interfaces, the
reduction of the potential for typing errors is probably EXECUTING THE JOB
their greatest advantage.
There are two different ways to directly execute jobs
The last feature of mapping steps is the specification entered into SAS/Warehouse Administrator. From
of filtering on the WHERE tab. WHERE expressions the Job Hierarchy pane in the Process Editor, right-
may also be built in a “point and click” fashion in the click on the job to be run and select Run… A dialog
Expression Builder. Use the WHERE tab to specify box similar to Figure 13 will appear.
the merge keys when constructing a join process
and to set filtering when sub-setting a table.

LOAD STEPS

Load steps are where one can place user-written

code to override WA-generated SAS load code. To
specify a load step, right-click on a output table or
file in the Process View pane in the Process Editor
and select Edit Load Step. The dialog box similar to
the one illustrated in Figure 12 should appear.
Figure 13

Click on the Edit button to generate the SAS code

for the job in a Preview window and to edit it before
submitting it. This is similar to selecting
“View Code >” when right-clicking on a job.

Click on the Save button to generate the code

associated with a job and save it to a catalog source
entry or external file. This feature is very useful
when the job is large and static. Once generated,
Figure 12 an external job scheduler can launch the job. The
Submit button causes a job to be generated and
Select “User Written” and specify the catalog source executed directly.
entry that contains the load code. Click on the Edit
button to create or modify the load step. To schedule a job through SAS/Warehouse
Administrator, it is necessary to set up a scheduling
While there are many valid occasions where a user- server as part of the global metadata and a job
written load step must be specified, there is a information library. If only a single host computer is
temptation to supply user-written load steps when used to run jobs, an ordinary libref allocation for the

8
job information library will suffice. However, if http://www.sas.com/rnd/warehousing/wa/addins.html
multiple hosts run jobs, then SAS/SHARE should be
used. The list of add-in tools changes periodically and new
versions of existing tools are often available for
SCHEDULING JOBS download. The application interface to
SAS/Warehouse Administrator is documented so
SAS/Warehouse Administrator can natively send one can create their own add-in tools if they can
jobs to CRON (Unix hosts) and AT (Windows hosts). code in SAS Component Language (SCL).
The null scheduler generates a “stub” file that
external job schedulers can read for scheduling It is the author’s understanding the add-in tools will
information. To use the null scheduler, right-click on disappear in a future version of SAS/Warehouse
a job and select Properties. On the Date/Time tab, Administrator although the functionality that they
as illustrated by Figure 14, select when the job is to provide should remain.
be run. Then on the Server tab, specify the null
scheduler server.
GENERATING HTML DOCUMENTATION

One of the author’s favorite add-in tools is the one

that automatically generates HTML documentation.
To bring up this particular add-in, get out of the
Process Editor and select Tools -> Add-Ins ->
Publish metadata to HTML page from the pull-down
menu. A dialog box similar to the one illustrated in
Figure 15 should appear.

Figure 14

One external scheduler that takes advantage of the

null scheduler feature via an add-in tool is the LSF
JobScheduler. LSF JobScheduler is a product of
Platform Computing. An OEM license for LSF
JobScheduler is supplied as part of SAS/Warehouse
Administrator. To use LSF JobScheduler, it is Figure 15
necessary to request software keys from Platform
Computing. LSF JobScheduler should be The resulting HTML generates a header and table of
considered when dependent job scheduling or load- contents similar to what is shown in Figure 16.
sharing is desired.

ADD-IN TOOLS

Add-in tools are programs written by the SAS

warehouse developers (or users) to extend the
functionality of SAS/Warehouse Administrator. They
are installed on top of SAS/Warehouse
Administrator and used to help load external data,
model processes, schedule jobs, and to analyze,
search, and report on metadata.

Add-in tools are usually accessed by right-clicking

on an item in the Process Editor and selecting Add-
Ins… The link to the information on add-in tools on Figure 16
the SAS web site is:

9
SEARCHING AND MIGRATING METADATA Java interface so it will no longer be necessary to be
sitting in front of the host computer or to operate it
One of the big advantages of entering all of the via terminal emulation software. The Java interface
warehouse details as metadata is that one can will communicate to a metadata repository and
search it. From the pull-down menu, select Tools -> server.
Search Metadata… The dialog box similar to
Figure 17 should appear. There will be a one-way conversion tool to migrate
SAS/Warehouse Administrator metadata into Data
Builder. However, existing SAS/Warehouse
Administrator users can continue to use the product
as in the past.

CONCLUSION

The author hopes this paper has explained his

passion for using SAS/Warehouse Administrator
over traditional methods for creating and maintaining
data warehouses and marts. He also hopes that this
paper clearly illustrated how data warehouses are
Figure 17 modeled in SAS/Warehouse Administrator and
highlighted how metadata is created and managed.
Enter the search string. Because the search is Last, the author hopes that the tips passed by this
conducted on metadata, it can be restricted by paper will reduce the learning curve by other users
warehouse element type. To go directly to an item of this product.
shown in the results window, just double-click on the
Warehouse Element Type.
ACKNOWLEDGEMENTS
Another useful tool is the Metadata Copy wizard. If
one needs to move metadata to a different directory AppDev Studio, SAS, SAS/ACCESS, SAS/SHARE,
path, go to the SAS/Warehouse Administrator SAS/Warehouse Administrator, and WebHound are
desktop. Right-click on the warehouse environment trademarks of SAS Institute Inc. Microsoft, SQL
to be copied and select Copy… Follow the Server, and Microsoft Windows are trademarks of
instructions given by the wizard. the Microsoft Corporation. Oracle is a registered
trademark of Oracle Corporation.
One of the author’s tricks when modeling
warehouses on his laptop is to map a project to the The author would like to thank the Hartford Area
same drive letter and path as is used on the client’s SAS User Group Steering Committee, which
host computer. To migrate the warehouse to host encouraged him to prepare this paper. Special
computer, he merely copies the metadata physical thanks also go to Jon Schiltz and Tina Hobbs, SAS
directory to a CD-ROM or Zip disk and then copies it Technical Support Department, and to the author’s
to the host computer. colleagues at The Nash Engineering Company and
Pfizer Inc.

FUTURE CHANGES AND ENHANCEMENTS

CONTACT INFORMATION
At SUGI 27, the author visited with some of SAS
staff responsible for future versions of The author may be contacted as follows:
SAS/Warehouse Administrator. Among some of the
improvements anticipated for future releases were: Michael L. Davis
Bassett Consulting Services, Inc.
• Multiple-table join tools 10 Pleasant Drive
North Haven CT 06473-3712
• Enhancements to take advantage of
E-Mail: michael@bassettconsulting.com
multi-threading in SAS Version 9
Web: http://www.bassettconsulting.com
• Integration of the File Import Wizard Telephone: (203) 562-0640
Facsimile: (203) 498-1414
Also on the horizon was a new version of the
product called Data Builder. Data Builder provides a

10