Zip items from a start and end item list following inclusion/exclusion rules so as to correspond to the formatted categories of a single label.

%let cat=%label_zip(start, end, sexcl=, eexcl=, type=, sep=%str( ));

Arguments

start : (list of) starting item(s), i.e. the lower bound of values to consider in the categories of a label;
end : (list of) ending item(s), i.e. the upper bound of category; must be the same length as start above; the items in end are possibly equals to those in start when a single value is to be considered;
sexcl : (option) list of boolean flag(s) (Y/N) set to exclude items from the starting list; when the item in i-th position in sexcl is set to Y, then the item in the same position in start will be excluded from the label (i.e. regarded as lower bound); when set, must be the same length as start above; default: not set, as if it was set to N for all items in start;
eexcl : (option) ibid for excluding items of the ending list from the categories associated to the label; default: not set, as if it was set to N for all items in end;
type : (option) flag set to the type of the considered lists, i.e. either numeric (N) or char (C); default: 'N', i.e. the lists are processed as if they were of NUMERIC type;
sep : (option) character/string separator in input list; default: %str( ).

Returns

cat : concatenated list (i.e., comma-separated whenever the length of input start and end lists is >1) of items zipped from start and end; when considering lists of length 1, say it otherwise: start=s and end=e, then the zipping rule writes:

s if s=e, or
s-e if s<>e and sexcl=N, eexcl=N, or
s<-e if s<>e and sexcl=Y, eexcl=N, or
s-<e if s<>e and sexcl=N, eexcl=Y, or
s<-<e if s<>e and sexcl=Y, eexcl=Y,

if the label if of type 'N'. This can be easily extended for labels of type 'C', and extrapolated for lists of length >1 so as to be read like a category label.

Examples

Given two NUMERIC start/end lists and the corresponding exclusion rules:

%let start= 1  2  4     3  8;
%let end=   1  2  HIGH  5  10;
%let sexcl= N  N  Y     Y  N;
%let eexcl= N  N  N     Y  N;

it is possible to retrieve the categorisation of the derived label using:

%let cat=%label_zip(&start, &end, sexcl=&sexcl, eexcl=&eexcl);

which returns cat=1, 2, 4<-HIGH, 3<-<5, 8-10. In the case of a CHAR lists, we can run:

%let start=000 OTHER 1 A B;
%let end=&start;
%let cat=%label_zip(&start, &end, type=C);      

which returns cat="000", "OTHER", "1", "A", "B".

Run macro %_example_label_zip for more examples.

Notes

This can be used to define the unique categorization associated to a single label. Say it otherwise, given the format myformat below, with one label A only, stored in a table TMP:

PROC FORMAT library=WORK.formats cntlout=TMP;   
    VALUE myformat 
        1-3, 5, 10<-HIGH = "A";
run;

It is then possible to retrieve the categorisation of A since the following commands:

%let start=; %let end=; %let sexcl=; %let eexcl;
%var_to_list(TMP, START, _varlst_=start);
%var_to_list(TMP, END, _varlst_=end);
%var_to_list(TMP, SEXCL, _varlst_=sexcl);
%var_to_list(TMP, EEXCL, _varlst_=eexcl);
%let list=%label_zip(&start, &end, sexcl=&sexcl, eexcl=&eexcl);

will return list=1-3, 5, 10<-HIGH.

Note the following possible use of %list_ based macros:

%let labels=%list_append(
         %list_append(
             %list_sequence(start=1, end=36, step=5), 
             %list_sequence(start=5, end=40, step=5), 
             zip=%str(-)
             ), 
         %list_quote(
             %list_append(
                 %list_ones(%list_length(&start), item=Y), 
                 %list_append(&start, &end, zip=%str(_)),
                 zip=_EMPTY_
                 ),
             rep=_BLANK_
             ), 
         zip=%str(=)
         );

which returns labels=1-5="Y1_5" 6-10="Y6_10" 11-15="Y11_15" 16-20="Y16_20" 21-25="Y21_25" 26-30="Y26_30" 31-35="Y31_35" 36-40="Y36_40".

References

Smiley, C.A. (1999): "A fast format macro � How to quickly create a format by specifying the endpoints".

Arguments

Returns

Examples

Notes

References

See also