PROC OPTMODEL provides two primary means of producing formatted output. The PUT statement provides output of data values with detailed format control. The PRINT statement handles arrays and produces formatted output in tabular form.
The PUT statement is similar in syntax to the PUT statement in the DATA step and in PROC IML. The PUT statement can output data to the SAS log, the SAS listing, or an external file. Arguments to the PUT statement specify the data to output and provide instructions for formatting. The PUT statement provides enough control to create reports within PROC OPTMODEL. However, typically the PUT statement is used to produce output for debugging or to quickly check data values.
The following example demonstrates some features of the PUT statement:
proc optmodel; number a=1.7, b=2.8; set s={a,b}; put a b; /* list output */ put a= b=; /* named output */ put 'Value A: ' a 8.1 @30 'Value B: ' b 8.; /* formatted */ string str='Ratio (A/B) is:'; put str (a/b); /* strings and expressions */ put s=; /* named set output */
These statements produce the output in Figure 5.43.
Figure 5.43: PUT Statement Output
The first PUT statement demonstrates list output. The numeric data values are output in a default format, BEST12., with leading and trailing blanks removed. A blank space is inserted after each data value is output. The second PUT statement uses the equal sign (=) to request that the variable name be output along with the regular list output.
The third PUT statement demonstrates formatted output. It uses the @ operator to position the output in a specific column.
This style of output can be used in report generation. The format specification "8." causes the displayed value of parameter
b
to be rounded.
The fourth PUT statement shows the output of a string value, str
. It also outputs the value of an expression enclosed in parentheses. The final PUT statement outputs a set along with its
name.
The default destination for PUT statement output is the SAS log. The FILE and CLOSEFILE statements can be used to send output to the SAS listing or to an external data file. Multiple files can be open at the same time. The FILE statement selects the current destination for PUT statement output, and the CLOSEFILE statement closes the corresponding file. See the section FILE Statement for more details.
The PRINT statement is designed to output numeric and string data in the form of tables. The PRINT statement handles the details of formatting automatically. However, the output format can be overridden by PROC OPTMODEL options and through Output Delivery System (ODS) facilities.
The PRINT statement can output array data in a table form that contains a row for each combination of array index values. This form uses columns to display the array index values for each row and uses other columns to display the value of each requested data item. The following statements demonstrate the table form:
proc optmodel; number square{i in 0..5} = i*i; number recip{i in 1..5} = 1/i; print square recip;
The PRINT statement produces the output in Figure 5.44.
Figure 5.44: PRINT Statement Output (List Form)
The first table column, labeled "[1]," contains the index values for the parameters square
and recip
. The columns that are labeled "square" and "recip" contain the parameter values for each array index. For example, the last
row corresponds to the index 5 and the value in the last column is 0.2, which is the value of recip[5]
.
Note that the first row of the table contains no value in the recip
column. Parameter location recip[0]
does not have a valid index, so no value is printed. The PRINT statement does not display variables that are undefined or
have invalid indices. This permits arrays that have similar indexing to be printed together. The sets of defined indices in
the arrays are combined to generate the set of indices shown in the table.
Also note that the PRINT statement has assigned formats and widths that differ between the square
and recip
columns. The PRINT statement assigns a default fixed-point format to produce the best overall output for each data column.
The format that is selected depends on the PDIGITS=
and PWIDTH=
options.
The PDIGITS= and PWIDTH= options specify the desired significant digits and formatted width, respectively. If the range of
magnitudes is large enough that no suitable format can be found, then the data item is displayed in scientific format. The
table in the preceding example displays the last column with five decimal places in order to display the five significant
digits that were requested by the default PDIGITS=
value. The square
column, on the other hand, does not need any decimal places.
The PRINT statement can also display two-dimensional arrays in matrix form. If the list following the PRINT statement contains only a single array that has two index elements, then the array is displayed in matrix form when it is sufficiently dense (otherwise the display is in table form). In this form the left-most column contains the values of the first index element. The remaining columns correspond to and are labeled by the values of the second index element. The following statements print an example of matrix form:
proc optmodel; set R=1..6; set C=1..4; number a{i in R, j in C} = 10*i+j; print a;
The PRINT statement produces the output in Figure 5.45.
Figure 5.45: PRINT Statement Output (Matrix Form)
In the example the first index element ranges from 1 to 6 and corresponds to the table rows. The second index element ranges
from 1 to 4 and corresponds to the table columns. Array values can be found based on the row and column values. For example,
the value of parameter a[3,2]
is 32. This location is found in the table in the row labeled "3" and the column labeled "2."