PQL Procedures

Home	Starting	Environment	DBMS	VisualPQL	Procedures	SQL	Forms	Host/API	Index

PQL Procedures

Condescriptive

CONDESCRIPTIVE produces statistics for numeric variables in the procedure table. It is usually used with continuous variables that can assume any value in a given range. Discrete variables are usually analysed with FREQUENCIES.

CONDESCRIPTIVE may not be used with string variables. If a string variable is specified, a warning message is issued and the variable is ignored.

CONDESCRIPTIVE    [FILENAME = filename ]
         [ VARIABLES = varname [AS varname] .... | ALL ]
         [ WEIGHT  = varname]
         [ SAMPLE  = fraction]
         [ BOOLEAN = (logical expression)]
         [ TITLE   = 'text' ]
         [ STATISTICS = keyword list ]
         [ NOLABELS ]

FILENAME

Specifies the filename created by the procedure.
If multiple CONDESCRIPTIVE procedures are specified and this clause is omitted, output is appended to the end of the previous output file.
If no filename is specified, the output is written to the default output file (the scrolled output buffer in an interactive session).

VARIABLES

Specifies the procedure table variables for which statistics are produced. The order in which they are specified is the order in which they appear in the output file. If this option is not specified the default variables are output. Specify the

AS
varname

option to rename variables on output. This is particularly useful for array elements, where otherwise just the array name is used. Specify the AS option separately for each variable to rename:

VARIABLES = S(1) AS SALARY, S(2) AS TAX

WEIGHT

Specifies the procedure variable used to weight the variables. A weight of 1 is the default. The WEIGHT affects all statistics except MINIMUM and MAXIMUM values.

BOOLEAN

Specifies which procedure table records are used by the procedure. The procedure table records for which the logical expression is true are used by the procedure. If this option is not specified, all procedure table records are used.

SAMPLE

Specifies that a random sample of the procedure table records are used by the procedure.
The fraction specifies the percent of records used and is specified as a positive decimal number less than or equal to 1 (one). .25, for example specifies that a 25% sample be used.

TITLE

The TITLE clause defines the subtitle that is printed on the CONDESCRIPTIVE report. The subtitle text must be enclosed in quotes. If the TITLE clause is not used, the subtitle line is left blank.

STATISTICS

The STATISTICS clause contains a list of keywords that name the statistics produced by CONDESCRIPTIVE. ALL is the default. The keywords are:

`ALL`	produces all the statistics. Has the same result as omitting the `STATISTICS` clause or specifying all of the keywords.
`WCOUNT`	weighted count of non-missing values
`MAX`	maximum value
`MIN`	minimum value
`MEAN`	mean value
`STDV`	standard deviation
`SKEW`	skewness
`KURT`	kurtosis
`VAR`	variance
`STDE`	standard error of the mean
`CV`	coefficient of variability
`CI`	95 percent confidence interval about the mean
`SUM`	sum

NOLABELS Specifies that variable labels are not printed. By default, the CONDESCRIPTIVE report lists both the variable name and label.

Example

In the following simple example, the procedure table contains one record for each record type 1 record in the database. Statistics are then produced by the condescriptive procedure.

RETRIEVAL
PROCESS CASES
.  PROCESS REC 1
.    GET VARS SALARY
.    PERFORM PROCS
.  END REC
END CASE
CONDESCRIPTIVE VARIABLES =SALARY /
               FILENAME  = COND.TXT
END RETRIEVAL

The output is:

SALARY    Current monthly salary

WGT CNT           20.000            MAXIMUM                   4000.000
MINIMUM         1650.000            MEAN                      2780.000
STD              573.631            SKEWNESS                      .217
KURTOS             -.161            VARIANCE                329052.632
STD ERR          128.268            C.V. PCT                    20.634
.95 C.I.        2511.532            TO                        3048.468
SUM            55600.000

VALID OBSERVATIONS                          20
MISSING OBSERVATIONS                        0