How to drop variables in a do

[Mathis91]

Hello, I want to drop a variable in a dataset which came from a transpose.

The difficulty here is the number of variables to drop : it can change following the number of columns at the end of the transpose.

I've tried this : 

data datasetname;
set datasetname;
array date[&Dimension_tbl.];

do i=1 to &Dimension_tbl.;
drop date{i};
end;
run;

I don't understand why it doesn't work. Of course, the &Dimension_tbl. is defined earlier.

[PeterClemmensen]

A few things. 

1. The Drop Statement is a complie-time only statement. Meaning that it is not considered during execution of the data step. Therefore it cannot be called in conditional logic.
2. What does &Dimension_tbl. contain?
3. Since your data is from a Transpose, I assume that the variables (columns) you want to drop share a common prefix, correct?

[Mathis91]

Thanks for your reply.

&Dimension_tbl. contains the number of columns created during the transpose, it's the number of date variables

And yes, the variables that I want to drop share a common prefix, which is "date". The suffix is between 1 and &Dimension_tbl.

[PeterClemmensen · Accepted solution]

Ok. Then you can use the logic by @LinusH or simply do

drop date:;

Which will drop all variables prefixed with 'date' .

[Mathis91]

Thanks a lot, I didn't know about this feature !

[LinusH]

No, but you can use a variable range for this:

drop date1-date&Dimension_tbl.;

https://documentation.sas.com/doc/en/pgmsascdc/9.4_3.5/lrcon/p0wphcpsfgx6o7n1sjtqzizp1n39.htm#:~:tex...

[Mathis91]

Thanks but it didn't work.
In the documentation I didn't see an example with a macrovariable, I don't know if it's me who doesn't understand how to use it

[Tom]

---

@LinusH wrote:

No, but you can use a variable range for this:

drop date1-date&Dimension_tbl.;

https://documentation.sas.com/doc/en/pgmsascdc/9.4_3.5/lrcon/p0wphcpsfgx6o7n1sjtqzizp1n39.htm#:~:tex...

---

This is the better answer than use date: variable list.  If you drop using the date prefix it might drop other variables that were not intended.