Oh, I forgot to mention that my original dataset contained only one observation. If there had been many more the log would have grown very long & repetitive! Adding STOP; as the last statement in the data _null_ step would help, I think.

JW

On Thu, 26 Apr 2001 10:47:34 -0400, Jay Weedon <jweedon@earthlink.net> wrote:

>Hi folks, > >I know this type of query comes up here every now & then, and the >solutions I've generally seen involve writing somewhat complex macros. > >I'm sure I'm not the first to stumble over this alternative, but FWIW >here it is. It's suitable for the situation where the new names >resemble the old names except for the last few characters. > >I had a dataset containing hundreds of numeric variables, the names of >which all of which end in the digit "3". My problem was to rename all >these 3s to 1s. > >I've taken the macro route before, but it always seems to take me >quite some time to get it working properly. This time I instead wrote >the following little program: > >data _null_; > set oldset; > array allvar _numeric_; > do over allvar; put allvar= 1. +(-2) allvar= 1. +(-4) '1 '; end; >run; > >and then simply pasting the resulting text from the log into the place >I've marked with an asterisk in a 2nd simple program > >data newset; > set oldset; > rename > * > ; >run; > >Jay Weedon.