var_check

• Check that a field/variable (defined as a string) actually exists in a given dataset.

%let ans=%var_check(dsn, var, lib=WORK);

• Trim a list of string elements to keep only those that actually define variables in a given dataset.

%var_check(dsn, var, _varlst_=, lib=WORK);

Arguments

• dsn : an input reference dataset;
• var : either a (list of) variable name(s), or the (list of) position(s) of variable(s) whose existence in input dataset dsn is tested; when var is passed as a list of integers, it is only verified that these values are in the range [1,#{variables in dsn}] where #{variables in dsn} is the number of dimensions in dsn; the list of variables of corresponding variables is then returned through _varlst_ (see below);
• lib : (option) output library; default (not passed or ' '): lib is set to WORK.

Returns

• ans : (option) the error code of the test, i.e.:

  • 0 when the variable var exists in the dataset,
  • 1 (error: "var does not exist") otherwise;

  should not be used contemporaneously with the option _varlst_ below; this is further unusable when var is passed as a list of integers;

• _varlst_ : (option) name of the macro variable used to store the string elements from var that do actually define existing variables/fields in dsn; incompatible with returned result ans above.

Examples

Let us consider the table _dstest5:

f	e	d	c	b	a
.	1	2	3	.	5

then:

%let var=a b y z c;
%let list=;
%let ans=%var_check(_dstest5, &var);

returns ans=0 0 1 1 0, the outputs of the existence test of string elements of var in _dstest5, while:

%var_check(_dstest5, &var, _varlst_=list);

returns list=a b c, the only string elements of var which define variables in _dstest5. Finally, it is possible to pass integer var to the macro so as to retrieve the names of the variables in corresponding positions, e.g.:

%let var=3 1 4 2;
%let list=;
%var_check(_dstest35, &var, _varlst_=list);

returns list=EQ_INC20 geo RB050a time.

Run macro %_example_var_check for more examples.

Notes

1. As mentioned above, two types of outputs are possible: either the answer ans to the test when a result shall be returned (and _varlst_ is not passed), or an updated list of acceptable variables (when _varlst_ is passed). The former case is useful when testing a single variable existence in a dataset, the latter for triming a list of actual variables. Contemporaneous use is impossible.
2. In short, the macro performs either of the two following operations:
   • the "intersection" between the list var with the list of variables in dsn, i.e. when _varlst_ is set:

%let allvars=;
%ds_contents(&dsn, _varlst_=allvars, lib=&lib);
%let &_varlst_=%list_intersection(&var,  &allvars, casense=NO);

• the test of existence of the variables var otherwise, i.e. when a single variable var is passed:

%let err=%sysfunc(varnum(%sysfunc(open(&lib..&dsn)),&var));
%if &err>0 %then    %let ans=0;
%else               %let ans=1;

1. The order of the variables in the output list &_varlst_ /answer ans matches that in the input list var.
2. When none of the string elements in var matches a variable in dsn, an empty list &_varlst_/ answer ans is set.

References

1. SAS community: "Tips: Check if a variable exists in a dataset".
2. Johnson, J. (2010): "OBJECT EXIST: A macro to check if a specified object exists".

See also

%ds_check, %lib_check, %ds_contents, %dir_check, %file_check, VARNUM.