Dropping variables with wildcard in Data Step

[deleted_user]

In Stata, if I want to drop variables before merging that all have an "h1" root, I would type:

drop h1*

Stata drops all variables beginning with "h1", no matter what order they are in.

In SAS, I try:

data temp (drop=h1*);

SAS does not recognize the wildcard. What is the proper way to do this?

Thanks

Paul

[sbb]

Use the colon character with SAS to specify a variable root / prefix.

Scott Barry
SBBWorks, Inc.

[deleted_user]

Thanks. Is there an example of this in the documentation somewhere?

P

[LinusH]

(drop=h1:)

/Linus

[deleted_user]

Thanks. The only things I could find about wildcards in the documentation referred to "*" and "%".

P

[deleted_user]

Hi,

it looks very easy, but...

Is there any way to drop the variables ENDING with some pattern? (Or more: variables that contain some pattern).

I've tried with

DROP=:h1

but it doesn't work. I could manage that using PROC CONTENTS and macrovariables, but I hope to find some more clear solution. Is there any?

Thanks
Jaroslav

[LinusH]

There is no direct support for this in data step syntax (to my knowledge). But it doesn't have to be complicated. Try using sql with dictionary.columns to create a variable list using INTO and SEPARATED BY. See an example from this SUGI paper:

http://www2.sas.com/proceedings/sugi30/057-30.pdf

Regards,
Linus

[Patrick]

Have a look at this thread. It might give you an idea how to query the dictionary table and find vars with a certain pattern.
HTH
Patrick

Ooops... Somehow the link got "lost": http://support.sas.com/forums/thread.jspa?messageID=14275㟃 Message was edited by: Patrick

[darrylovia]

In the online documentation go to SAS Variables; SAS Variable Lists