Manual data selection

   

Syntax

1. BD>select : [TYPE] D [OP R] [, ...]

2. BD>select :

where TYPE is one of the three symbols (+,-,&); OP is one of the symbols (=, >, <, , , <>); R is a real number; and D is the name of a data-carrier

The SELECT:  command is used to select and deselect data cases. For most of [B/D]'s commands, the data selection is determined automatically to be the largest selection appropriate to the situation. However, it may sometimes be useful or necessary to take some subset of this largest possible selection, and thus enable the manual selection of data. This is achieved by judicious use of the SELECT:  command, and by switching off the autoselect  control.

We assume that prior to this command being issued there exists a prior selection, P. You might think of a selection as a notional set of flags attached to each observation number, where the flag is switched on if the observation number is to be selected, and off otherwise. The second form of the syntax sets the prior selection to be no cases.

Each usage of the SELECT:  command consists of the following, repeated for as many data-carriers as you wish.

1. The name of a data-carrier, D.
2. Optionally, the TYPE of selection. Where no TYPE is given, the symbol `+' is assumed by default. The three TYPEs are as follows.

   -

   meaning that the selection is to become P and not D.

   +

   meaning that the selection is to become P or D.

   &

   meaning that the selection is to become P and D.

3. A qualifier, consisting of an operator OP and a real number R, which restrict the number of cases to those cases of D which satisfy the qualifier. The possible OPs are

   =

   meaning only those cases i for which ;

   <>

   meaning only those cases i for which ;

   <

   meaning only those cases i for which ;

   

   meaning only those cases i for which ;

   >

   meaning only those cases i for which ;

   

   meaning only those cases i for which .

• Examples