Database Creation

Table: Employee
Column Name
Employee_ID (PK)
Employee_Name
Manager_ID

Data type
AutoNumber
Text
Number

Column Name
Product_ID (PK)
Brand_Name
Product_Name
Price
Subcategory_ID

Data type
Number
Text
Text
Currency
Number

Column Name
Category_ID (PK)
Category

Data type
AutoNumber
Text

Table: Product

Table: Product Category

Table: Product Subcategory

Column Name
SubCategory_ID (PK)
Subcategory
Category_Manager
Category_ID

Data type
AutoNumber
Text
Text
Number

Table: SalesFact
Column Name
Month (PK)
State_ID (PK)
Product_ID (PK)
Employee_ID (PK)
Sales_Dollars
Sales_Units

Data type
Date/Time
Number
Number
Number
Currency
Number

Column Name
State_ID (PK)
State_Name
Region
Country

Data type
Number
Text
Text
Text

Table: State

Note: Insert at least 10 records in each table.

Practical No: 02
Title: Creating a Data Warehouse
Problem Statement: Create and schedule a DTS Package using Data Transformation
services (DTS) tool. Fire at least 5 queries on the database.
Description:
Creating a DTS Package with the DTS Import/Export Wizard
The Data Transformation Services (DTS) Import/Export Wizard offers the simplest
method of building a DTS package, interactively guiding you through the process of
copying and transforming data. Following are the basic steps for creating a package with
the DTS Import/Export Wizard:
1. Open SQL Server Enterprise Manager, expand the server, and click the Databases
folder.
2. On the Tools menu, point to Data Transformation Services, and then click
Import Data.
3. In the Choose a Data Source dialog box, select Microsoft Access as the Data
Source, and then type the path and file name of your Access database (.mdb), or
use the browser to browse for the file.
4. In the Choose a Destination dialog box, select Microsoft OLE DB Provider for
SQL Server, and then select the database server and the required authentication
mode. In the Database dialog box, select the target database on the server.
5. In the Specify Table Copy or Query dialog box, click Copy tables and views
from the source database. If you copy the data or the results of a query, you can
customize the data being copied to the destination. You can:

6. In the

Select which source or destination columns to copy.

Select which source or destination columns to ignore.
Change the data type where valid.
Define how the data is to be converted between source and destination.
Select Source Tables and Views dialog box, select the tables and queries

to import.
7. In the Save, Schedule, and Replicate Package dialog box, select Run
Immediately.(You can save the package to the SQL Server msdb database, to
SQL Server 2000 Meta Data Services, as a structured storage file (.dts file), or as
a Microsoft Visual Basic file.)
8. Click Finish.

Cube Creation
Create a new OLAP database
1. Right click the server in the console tree, and click New Database to display the
Database dialog Box.
2. Type Sales as the database name, and type Market Database as the description.
3. Click OK to close the dialog box and to create the database.
Specify a data source
1. Expand the Sales database folder. Right-click the Data Sources folder, and on the
shortcut menu, click New Data Source.
2. Select Microsoft Jet 4.0 OLE DB Provider, and click Next.
3. On the connection Tab of the Data Link Properties dialog box, click the ellipsis
button(), navigate to the folder containing the Sales.mdb
4. Click Test Connection to make sure the data source definition is correct. Then
click OK to close the dialog box. Expand the Data Sources folder to see the New
Data Source.
5. Right-click the New Data Source, and click Copy on the shortcut menu. Rightclick the Sales database, and click Paste on the shortcut menu. When asked for a
new name, click Delete on the shortcut menu, and click Yes to confirm.
Designing a Cube by Using the Cube Wizard
Select the Fact Tables and the Measures:
1. Right Click the Cube folder under the Sales database, point to New Cube, and
then click Wizard to start the Cube Wizard. On the Welcome screen, click Next.
2. Select SalesFact from the list of tables.
3. Click Browse Data to review the records in the table. Close the Browse Data
window.
4. After selecting the fact table, click Next to move to the next step of the wizard.
5. Add the Sales_Dollars and Sales_Units columns as measures by double-clicking
each in turn. You can select a measure and click the right arrow button(>).
6. After adding both measures, click Next.

Create a dimension from a Star Schema table:

1. Click a New Dimension in the Wizard to start the Dimension Wizard. Select the
option to skip the welcome screen, and then click on Next.
2. Click the Option Star Schema: A Single Dimension Table, and the click Next.
3. In the Select The Dimension Table Screen, select State from the Available
Tables list.
4. After selecting the dimension table, click Next.
5. Double-click Country, Region, and State_Name in the order. You select the
Levels from the most summarized to the most detailed.
6. Click Next three times. Type State as the name of the dimension. Before closing
the wizard you can preview the dimension hierarchy.
Create a dimension from Snowflake schema tables:
1. Click Next Dimension in the Cube wizard to again launch the Dimension Wizard.
2. Click the option Snowflake Schema: Multiple, Related Dimension Tables,
and Click Next.
3. Double-click the dimension tables in the following order: Product, Product
Subcategory, and Product Category.
4. Click Next.
5. Review the joins between tables. If necessary add join between the two
Subcategory_ID columns and the two Category_ID columns, and click Next.
6. Double-click Category, Product_Name, and Subcategory intentionally
reversing the order of the last two levels. The Dimension Wizard displays a
message suggesting that you can rearrange the levels.
7. Click Yes to put the Product_Name level below the Subcategory level.
8. Click Next three times.
9. Type Product as the name of the dimension, expand items in the Preview window
as desired, and click Finish.

Create a dimension from Parent-Child tables:

1. Click Next Dimension in the Cube wizard to again launch the Dimension Wizard.
2. Click Parent-Child: Two Related Columns In a Single Dimension Table, and click
Next.
3. Select Employee as the dimension table, and click Next.
4. In the Member Key drop-down list, select Employee_ID. This is the column that
is joined to the fact table. In the Parent Key drop-down list, select Manager_ID.
In the Member Name drop-down list, select Employee_Name.
5. Click Next twice: once after defining the parent-child hierarchy, and once to
bypass the advanced options step.
6. Type Employee as the name of the dimension.
7. Click Finish.
Create a Time dimension from a date column:
1. Click Next Dimension in the Cube wizard to again launch the Dimension Wizard.
2. Click the Option Star Schema: A Single Dimension Table, and the click Next.
3. In the Select The Dimension Table Screen, select SaleFact from the Available
Tables list. Because the SaleFact table contains a Date/Time column, the Wizard
displays a screen offering to create a Time dimension based on that column.
4. Select the dimension type, and then click the Time Dimension.
5. The Date column drop-down list allows you to choose which Date column to use
if there is more than one in the table. Select Month and click Next.
6. On the screen that asks you to create the Time dimension levels, you can select
the type of Time dimension hierarchy you want. You can even specify the Day
and Month for the Year to start useful for dealing with the fiscal years. In the
select time levels drop-down list, click Year,Quarter,Month and then click
Next.
7. When you use the Wizard to generate the date related levels you do not select
columns to use for the levels. Click Next to bypass the Advanced Option
Screen.
8. In the final screen on the Dimension Wizard, type Time as the dimension name.

Name, Save and Process a Cube:

1. In the Cube Wizard click Next, click Yes when asked if you want to count the fact
table rows.
2. Type Sales as the name of the Cube and click Finish.
3. Click Exit on the File menu to close the Cube Editor. Click No when asked if
you want to set storage option.
4. Expand the Cubes folder of the Sales database. Right-click the newly created
SalesCube and click Process.
5. Click OK to begin processing.

Working with Measures in the Cube

(To create a derived measure)
1. Right-click on SalesCube and select Edit.
2. In the SalesCube cube, right-click the Measures folder and click New Measure.
Select the Sales_Dollars column, and click OK. The measure is added with the
name Sales_Dollars1.
3. Change the Source Column property to SalesFact.Sales_Dollars*0.15, and
type Commission for the Name property.
4. On the Tools menu, click Process Cube, agree to save the cube, decline design
aggregates, and click OK when asked to specify the processing method. Close the
Process log window, and browse the data.
5. Scroll the grid as needed to see the new Commission measure.
(To specify the aggregation function for a measure)
1. In the Cube Editor, Click the Insert Measure toolbar button, and click doubleclick the Sales_Units column. This creates a measure named Sales_Units1. Type
Bad Price for the measures Name property, and change the value of the Siurce
Column property to SalesFact.Sales_Dollars:/SalesFact.Sales_Units. This
is a derived measure. It uses the default aggregation function: Sum.
2. Process the Cube that is, click the Process Cube button, save the cube, dont
design aggregates, and choose the full process method. Close the Process log
window, and browse the data.
3. Change the name of the Bad Price to Min Price, and change the value of the
Aggregate function to Min.
4. On the Tools menu, click Process Cube.
(To create a calculated measure)
1. In the Cube tree, right-click the Calculated Members folder, and click New
Calculated Member. In the Calculated Member dialog box, type Net Price as the
new member name.
2. In the data tree, fully expand the Measures dimension and double-click the Sales
Dollars measure to make the member name appear in the Value Expression box.
Click the slash ( / ) on the right side of the dialog box, and then double-click Sales
Units in the Data tree.
3. Click OK.

Creating an Excel Pivot Table and Pivot Chart

1. Open MS-Excel. Click on Data menu, click PivotTable And PivotChart Report.
2. Click External Data Source, and then click Next. Step 2 of the wizard allows you
to specify the external data source.
3. Click Get Data. Clicking the button starts Microsoft Query. In the Choose Data
Source dialog box, click the OLAP Cubes tab.
4. Select <New Data Source> and click OK.
5. Type Sales as the name of the data source, and select Microsoft OLE DB Provider
For OLAP Services 8.0 as the OLAP provider. Then click Connect.
6. In the Multidimensional Connection dialog box, leave Analysis Server selected as
the data source type, type the server computer name or localhost if you are
running Excel on the server machine in the Server box, click Next.
7. Select Sales as the database, and click Finish.
8. Click OK to continue. The Sales data source now appears in the list of OLAP
cubes.
9. With the Sales data source selected, click OK.
10. Click Finish to create a skeleton PivotTable Report, along with a PivotTable
toolbar that displays all the dimensions and measures in the cube. The PivotTable
toolbar with its buttons for the dimensions and measures is the key to creating
an Excel PivotTable report.
11. Drag the Sales Units button from the PivotTable toolbar onto the data area, which
is labeled Drop Data Items here. The PivotTable report changes from a skeletal
framework to an actual report, showing a single cell, the total sales units.
12. Drag the Product dimension button to the row axis of the PivotTable report,
which is cell A4 on the worksheet. The category names appear as row headings.
13. Drag the Time dimension button to the column axis of the PivotTable report,
which is cell B3 on the worksheet. The year labels appear as column headings.
14. Drag the State dimension button to the page axis of the PivotTable report, which
is cell A1 of the worksheet.
15. To view a Pivot Chart, on PivotTable toolbar click the Chart Wizard button.

Member Property Creation and new Dimension Creation

To create Member Properties:
1. In the Analysis Manager tree pane, Right Click on the Shared Dimensions folder.
2. Select Editor.
3. Select the Product dimension.
4. Right Click on the field Brand_Name in the Product Dimension and select
Insert As Level.
5. Right Click on the field Product_Name in the Product Dimension and select
Insert As Level.
6. Right Click on Product_Name level in the left-pane and select New Member
Property.
7. In the Insert Member Property dialog box select Price.
8. Click OK.
9. Click on File Save. Give new dimension name as Product_M.
10. Click on File Exit.
11. Right Click on SalesCube. Select Edit.
12. In the left-pane of the Cube Editor Right-Click on Dimensions folder. Select
Existing Dimensions. Select Product_M and press the button (>). Click OK.
13. Click on Save button. Click on Process button. Click No. Select Full Process.
Click on OK.
14. Click File Exit.

Firing queries on the Cube by using MDX Sample Application

1. Launch the MDX Sample Application, which is on the Analysis Services menu
along with Analysis Manager. Click OK to connect to the server. Click the New
Query File toolbar button to create a new query file. A query file can contain
multiple queries.
2. In the toolbar, select Sales as the database name. Then, in the Cube drop-down
lost box, select SalesCube as the cube-name.
Note: MDX Sample application is divided into three vertical panes: the top pane
is the query pane, where one builds an MDX query statement. The middle pane is
the metadata pane, where one inspects information in a selected cube. The bottom
pane is the results pane, which shows the output of an MDX query.
3. Replace the contents of the query pane with SELECT FROM SalesCube.
4. Click the Run Query button on the toolbar.
5. Similarly, the following queries can be tried out:
Query1:
SELECT [Product].[Category].Members ON COLUMNS FROM SalesCube
Query2:
SELECT {[Sales Dollars],[Sales Units]} ON COLUMNS FROM SalesCube
Query3:
SELECT Country.Members ON COLUMNS, Category.Members ON ROWS
From SalesCube
Query4:
SELECT Crossjoin(Country.Members,Category.Members) ON COLUMNS
From SalesCube
Query5:
SELECT [Measures].Members ON COLUMNS, [Product].
[Subcategory].Members ON ROWS From [SalesCube]

Creation of a Mining Model with Microsoft Decision Trees Technique

1. Right Click on SalesCube. Select New Mining Model.
2. Select the Data Mining Technique as Microsoft Decision Trees. Click on Next.
3. Select the Dimension as Product_M and Level as Product_Name.
4. In the Mining Model Wizard select A member property of the cased level as
Price. Click Next.
5. In the Select training data step Click Next.
6. In the Create a dimension and virtual cube (optional) step, enter New_Product
in the Dimension name box. Then, in the Virtual cube name box, enter Trained
Cube. Click Next.
7. In the final step, type Product Pattern Discovery in the Model name field.
Ensure that Save and Process now is selected. Click Finish.
8. Click on Close after finishing the processing.
9. OLAP Mining Model Editor will be opened. In the left-pane expand the
Measures folder.
10. Click on Save button.
11. Click on Process button.

Creation of a Mining Model with Clustering Technique

1. Right Click on SalesCube. Select New Mining Model.
2. Click on Next. Select Source Type as OLAP data. Click on Next. Select
SalesCube, click on Next.
3. Select the Data Mining Technique as Microsoft Clustering. Click on Next.
4. Select the Dimension as Product_M and Level as Product_Name.
5. In the Mining Model Wizard select A member property of the cased level as
Price. Click Next.
6. In the Select training data step Click Next.
7. In the final step, type Product Clustering in the Model name field. Ensure that
Save and Process now is selected. Click Finish.
8. Click on Close after finishing the processing.
9. OLAP Mining Model Editor will be opened. In the left-pane expand the
Measures folder.
10. Click on Save button.
11. Click on Process button.