Home > Commands > Database > Group Data

Group Data

The Group Data command lets you aggregate employee information to produce a file in which each record represents a group of employees having similar characteristics. The number of employees in each group is stored in the field COUNT.

Description may be any descriptive phrase, including spaces, under which to save this data grouping.

Input file is a database containing seriatim (person by person) information. This database will be used produce a second grouped database; the input file will not be affected.

Output file is a new database to write grouped data to. If the output file already exists, ProVal will ask if you want to replace it when you run the grouping.

Ordinarily, coded fields (such as Sex or Status) are used for grouping fields. These fields are suitable for grouping because they typically have only a few distinct values. If you select a continuous field (such as Salary or Date of Termination), the Break Pts button lights up to allow you to enter breakpoint values.

Choosing groups which will adequately reflect the nuances of the data, plan, and valuation methods is more of an art than a science. A few rules of thumb are:

Click the BrkPts… button to enter breakpoint values. The breakpoint values are used to transform continuous fields into coded intervals, such as every integral age from 20 to 85.

To enter breakpoints, enter a list of numbers or dates separated by spaces. For example, if you enter the breakpoints

20 50 85

the field (call it “x”) will be divided into the following intervals:

x < 20

20 <= x < 50

50 <= x < 85

85 <= x .

The breakpoint values specify the inclusive lower bound for each interval. A shorthand notation is available for entering breakpoints. The following expression is equivalent to “20 21 22 23 … 82 83 84 85”:

From 20 to 85, step 1

Breakpoints are optional; if you omit them, the raw numeric or date values will be counted. This might make sense, for example, if there were only a few different salaries or termination dates.

Data fields are the numeric and date fields to be written out to the grouped file. Coded fields used as grouping fields are automatically written out to the grouped file. In practice, it is common to select all data fields rather than spend anytime discerning which are necessary for the valuation or core projection.

Click the AvgMeths… button to choose between arithmetic and geometric averaging.

Enter Sex fields and associated Male codes to have ProVal calculate an average percent male (member) and contingent annuitant percent male for each cell of the grouped database (be sure that you don’t group on sex, or averaging of males and females will not take place). The fields PCTMALE and CAPCTMALE will be used to store this information.

Click the Selection expression topic to limit the number of seriatim records to be used in creating the grouped database. For example, you might use this to create a grouped database profiling new entrants hired during the last three (or any number of) years.

Click the Data Defaults topic to apply default values to input in database fields before grouping the data. This is done by specifying Census Specifications containing data defaults. (See Data Defaults for more information.) Select from the list of Census Specifications already defined in the current Project or, to create a new set of Census Specifications, or modify an existing one, click the image/backdoor_button.gif button. The Project selected (at the bottom of the dialog box) determines the Census Specifications choices in the list. If defaults were used in the Valuation you’re trying to replicate with grouped data, you should apply defaults here for best results. Note that if the option to default numeric field values to zero (except for salary fields) is selected, all numeric fields, including salary fields, will be given zero default values.