ds_copy

Create a working copy of a given dataset.

%ds_copy(idsn, odsn, where=, groupby=, having=, mirror=COPY, ilib=WORK, olib=WORK);

Arguments

• idsn : a dataset reference;
• groupby, where, having : (option) expressions used to refine the selection when mirror=COPY, like in a SELECT statement of PROC SQL (GROUP BY, WHERE, HAVING clauses); these options are therefore incompatible with mirror=LIKE; note that where and having should be passed with %quote; see also %ds_select; default: empty;
• mirror : (option) type of copy operation used for creating the working dataset, i.e. either an actual copy of the table (mirror=COPY) or simply a copy of its structure (i.e., the output table is shaped like the input ones, with same variables: mirror=LIKE); default: mirror=COPY;
• ilib : (option) name of the input library; by default: empty, i.e. WORK is used.

Returns

• odsn : name of the output dataset (in WORK library) where a copy of the original dataset or its structure is stored;
• olib : (option) name of the output library; by default: empty, i.e. WORK is used.

Example

For instance, we can run:

%ds_copy(idsn, odsn, mirror=COPY, where=%quote(var=1000));

so as to retrieve:

DATA WORK.&odsn;
    SET &ilib..&idsn;
    WHERE &var=1000; 
run; 

See %_example_ds_copy for more examples.

Note

The command ds_copy(idsn, odsn, mirror=COPY, ilib=ilib, olib=olib) consists in running:

DATA &olib..&odsn;
    SET &ilib..&idsn;
run; 

while the command ds_copy(idsn, odsn, mirror=LIKE, ilib=ilib, olib=olib) is equivalent to:

PROC SQL noprint;
    CREATE TABLE &olib..&odsn like &ilib..&idsn; 
quit; 

See also

%ds_select.