Language Reference

SUMMARY Statement

computes summary statistics for SAS data sets

SUMMARY <CLASS operand> <VAR operand> <WEIGHT operand>

           <STAT operand> <OPT operand> <WHERE(expression)>;

where the operands used by most clauses take either a matrix name, a matrix literal, or an expression yielding a matrix name or value. A discussion of the clauses and operands follows.

The SUMMARY statement computes statistics for numeric variables for an entire data set or a subset of observations in the data set. The statistics can be stratified by the use of CLASS variables. The computed statistics are displayed in tabular form and optionally can be saved in matrices. Like most other IML data processing statements, the SUMMARY statement works on the current data set.

The following options are available with the SUMMARY statement:

CLASS operand

specifies the variables in the current input SAS data set to be used to group the summaries. The operand is a character matrix containing the names of the variables, for example:

  
    summary class { age sex} ;
 

Both numeric and character variables can be used as CLASS variables.

VAR operand

calculates statistics for a set of numeric variables from the current input data set. The operand is a character matrix containing the names of the variables. Also, the special keyword _NUM_ can be used as a VAR operand to specify all numeric variables. If the VAR clause is missing, the SUMMARY statement produces only the number of observations in each class group.

WEIGHT operand

specifies a character value containing the name of a numeric variable in the current data set whose values are to be used to weight each observation. Only one variable can be specified.

STAT operand

computes the statistics specified. The operand is a character matrix containing the names of statistics. For example, to get the mean and standard deviation, specify the following:

  
    summary stat{mean std};
 

The following list describes the keywords that can be specified as the STAT operand:

CSS

computes the corrected sum of squares.

MAX

computes the maximum value.

MEAN

computes the mean.

MIN

computes the minimum value.

N

computes the number of observations in the subgroup used in the computation of the various statistics for the corresponding analysis variable.

NMISS

computes the number of observations in the subgroup having missing values for the analysis variable.

STD

computes the standard deviation.

SUM

computes the sum.

SUMWGT

computes the sum of the WEIGHT variable values if WEIGHT is specified; otherwise, IML computes the number of observations used in the computation of statistics.

USS

computes the uncorrected sum of squares.

VAR

computes the variance.

When the STAT clause is omitted, the SUMMARY statement computes these statistics for each variable in the VAR clause:

• MAX
• MEAN
• MIN
• STD

Note that NOBS, the number of observations in each CLASS group, is always given.

OPT operand

sets the PRINT or NOPRINT and SAVE or NOSAVE options. The NOPRINT option suppresses the printing of the results from the SUMMARY statement. The SAVE option requests that the SUMMARY statement save the resultant statistics in matrices. The operand is a character matrix containing one or more of the options.

When the SAVE option is set, the SUMMARY statement creates a CLASS vector for each CLASS variable, a statistic matrix for each analysis variable, and a column vector named _NOBS_. The CLASS vectors are named by the corresponding CLASS variable and have an equal number of rows. There are as many rows as there are subgroups defined by the interaction of all CLASS variables. The statistic matrices are named by the corresponding analysis variable. Each column of the statistic matrix corresponds to a statistic requested, and each row corresponds to the statistics of the subgroup defined by the CLASS variables. If no CLASS variable has been specified, each statistic matrix has one row, containing the statistics of the entire population. The _NOBS_ vector contains the number of observations for each subgroup.

The default is PRINT NOSAVE.

WHERE expression

conditionally selects observations, within the range specification, according to conditions given in expression. The general form of the WHERE clause is

WHERE( variable comparison-op operand)

In the preceding statement,

variable

is a variable in the SAS data set.

comparison-op

is one of the following comparison operators:

<

less than

<=

less than or equal to

=

equal to

>

greater than

>=

greater than or equal to

^=

not equal to

?

contains a given string

^?

does not contain a given string

=:

begins with a given string

=*

sounds like or is spelled like a given string

operand

is a literal value, a matrix name, or an expression in parentheses.

WHERE comparison arguments can be matrices. For the following operators, the WHERE clause succeeds if all the elements in the matrix satisfy the condition:

^=   ^?   <   <=   >   >=

For the following operators, the WHERE clause succeeds if any of the elements in the matrix satisfy the condition:

=   ?   =:   =*

Logical expressions can be specified within the WHERE clause, by using the AND (&) and OR (|) operators. The general form is

	clause&clause	(for an AND clause)
	clause|clause	(for an OR clause)

where clause can be a comparison, a parenthesized clause, or a logical expression clause that is evaluated by using operator precedence.

Note: The expression on the left-hand side refers to values of the data set variables, and the expression on the right-hand side refers to matrix values.

See Chapter 6 for an example that uses the SUMMARY statement.