Example: Modifying a Data Table
In this example, you combine
selected columns from two data sets and edit them in a new data
table. This example assumes that you have no data set loaded in
the Analyst data table. If you do, select File New
before starting the example.
Each data set contains the results of taste tests of breakfast
cereal. Each cereal is rated by several judges, on a scale of 1 to 5.
After you concatenate the two data sets, you split the rating
column by sample number.
To select the data sets and bring them
into a new Analyst data table, follow these steps:
- Select Tools Sample Data ...
- Select JRating1 and JRating2.
- Click OK to create the sample data sets in your Sasuser directory.
- Select Data Combine Tables
Concatenate By Rows ...
- Click on the Open SAS Data button. Select Sasuser
from the list of Libraries. Select Jrating1 from the
list of members. Click OK.
- In the Concatenate Tables by Rows dialog, click on the Open SAS Data
button again. Select Sasuser from the list of Libraries.
Select Jrating2 from the list of members. Click OK.
Figure 2.29: Concatenate Tables by Rows Dialog
- Select Interleave.
- Select JUDGE and SAMPLE from the list of Common
variables and click on the Interleave By button to use JUDGE
and SAMPLE as the variables by which the rows of the data tables will
be combined.
Figure 2.30: Interleave by Common Variables
- Click on the Variables button to select the columns to
include in the new data table.
Figure 2.31: Selected Columns for New Data Table
Only those columns common to both data tables
are kept by default, as shown in the Keep list. The column
SWEETNESS is not kept as part of the resulting table.
The number preceding the column name SWEETNESS
represents the data table to which this variable belongs.
- Click OK to return to the Concatenate Tables by Rows dialog.
Click OK again to display the new combined data table in a
results window.
Figure 2.32: Combined Table
- To modify the combined table, you need to open it in the
Analyst data table. Close the results window. Select the Combined
Table node in the project tree and click the right mouse button to
display the pop-up menu. Select Open.
Figure 2.33: Opening the Combined Table
- By default, data tables are opened in Browse mode. Select
Edit Mode Edit to
change the mode from Browse to Edit.
In the data table you can modify the data by splitting columns so that
a new column is generated when the value of a variable changes. You
can also subdivide data into ranges.
To subdivide the data into ranges and split the columns according to
sample number, follow these steps:
- Divide the taste test results into three
categories: good, mediocre, and bad. Select Data
Transform Recode Ranges ...
- Click on the arrow next to Column to recode: and select
Rating. Type taste_test in the New column name: field.
Change New column type: to Character. Type
3 in the Number of groups to be formed: field to designate three
taste test ranges.
Figure 2.34: Recode Ranges Information Dialog
Click OK to specify the new ranges.
- In the first row, type 0 in the Lower Bound column
and 2 in the Upper Bound column. Type bad in the
New Value column.
- When you press the Enter key, the upper bound
value of the previous row is automatically filled in as the lower
bound of the current row. Type 3 in the Upper Bound
column and mediocre in the New Value column.
- Move your cursor to the third row. Type 5 in the
Upper Bound column and good in the New Value column.
Figure 2.35: Boundary Values
- Click OK to save your new boundary values.
In the new table, the new ranges are displayed in the
taste_test column.
Figure 2.36: Table with taste_test Column
- Remove the Rating column by selecting the column and
selecting Delete ... from the pop-up menu. Click OK in the
Delete Items dialog.
Figure 2.37: Delete Rating Column
- You are going to split the taste_test column by the
Sample column so that a taste test for each sample is displayed by
judge. Select Data Split
Columns ...
- In the Split Columns dialog, select taste_test from the
list and click on the Split Column button. Select Sample from
the list and click on the Split By button.
- Select User-defined names for the column names. Type
Sample_ in the Column name prefix: field.
Figure 2.38: Taste_test Column Split by Sample
- Click OK. The resulting table displays the results of the
taste test by each participating judge.
Figure 2.39: Split Columns Table
Copyright © 1999 by SAS Institute Inc., Cary, NC, USA. All rights reserved.