Syntax
where C is an alphabetic character representing an index, and
where D is the name of a data-carrier.
The syntax for the INDEXVEC: command defines the name of an
index, C, and a range of values to be associated with it. To use this
syntax, it is necessary to have created a data-carrier D of length
k, with k data cases , beforehand. The
intention is to replace the character C successively by the integers
.
The command is used to vary the indices of quantities referenced in
COBUILD: and ITADJUST: commands. An index can be
overwritten - if an error occurs during re-definition, the former value
is retained. The values used for indices should round to positive
integers no larger than 32767. Index definitions can be deleted with the
XINDEX: command. When, for whatever reason, a data carrier is
deleted or overwritten, and if it is also
associated with an index, the index is deleted.
For example, suppose that you wish to construct (using the
COBUILD: command) the elements for ,
, and . Assuming that appropriate
functional beliefs have been specified, the following code example shows
how we could proceed.
index: j=1,1,4
data: <tempa>3
1 7 8
indexvec: i=tempa
data: <tempb>4
2 17 19 24
indexvec: k=tempb
cobuild: a.i.j.k
The example shows the irregular patterns being stored as the
data-carriers and respectively, and then
associated with indices respectively via the INDEXVEC:
command. We define the index j via the INDEX: command as it is
associated with a regular pattern.