FILE SELECT <new_data> FROM <old_data> BY <select_file>
with a specification KEYS=<var1>,<var2>,...
makes a <new_data> file by copying observations from the <old_data> file
selectively.
The observations to be copied are determined by <select_file> which also
is a Survo data file.
Only those observations which have same values in variables indicated by
the KEYS specification in <old_data> and <select_file> are copied to
<new_data>.
Options:
The cases that are copied from <old_data> can be marked in a variable
given by a MARK=<name_of_variable_in_old_data> specification.
The MARK variable must be created before FILE SELECT (and intialized
with 0 values). FILE SELECT then gives the value 1 for the MARK variable
in those cases which are copied.
By a specification NMATCH=<name_of_variable_in_select_file> the number
of cases matched by each key combination is counted and saved.
Also the NMATCH variable must exist before FILE SELECT (and initialized
with 0 values).
FILE SELECT is much faster than FILE COPY in applications where a subset
of cases is to be selected from a huge data file according to certain
identification keys.
A complementary selection of observations can be made by using a
KEYS!=<var1>,<var2>,... specification. Then cases which differ at least
in one of the key variables are selected.
D = More information on data management
Help System (web edition)