On 11/10/07, Howard Schreier <hs AT dc-sug DOT org> <nospam@howles.com> wrote: > > On Sat, 10 Nov 2007 11:17:34 -0000, Peter <crawfordsoftware@GMAIL.COM> > wrote: > > >On Nov 9, 9:04 pm, art...@NETSCAPE.NET (Arthur Tabachneck) wrote: > >> Excellent suggestion. However, to catch for potentially cancelling > >> negative and positive numbers, I'd change the statement to: > >> > >> if abs(min(of col:)) + abs(max(of col:)) ne 0; > >> > >> Art > >> -------- > >> > >> > >> > >> > >> > >> On Fri, 9 Nov 2007 14:49:11 -0600, data _null_, <datan...@GMAIL.COM> > wrote: > >> >Art, > >> > >> >Consider the following slight modification to your program. Using > >> >COL: allows the program to work with more than 4 observations > >> >(transposed to variables). > >> > >> >Would not work with lots of observations but the OP mentioned about > >> >100 as I recall. > >> > >> >data temp2; > >> > set temp1; > >> > if min(of col:) + max(of col:) ne 0; > >> > run; > >> > >> >On Nov 9, 2007 1:38 PM, Arthur Tabachneck <art...@netscape.net> wrote: > >> >> Mike, > >> > >> >> I don't know if this is any easier than Paige's suggestion, but it > >> avoids > >> >> having to deal with the nuances of macro variables: > >> > >> >> data have; > >> >> input a b c d e; > >> >> cards; > >> >> 1 0 2 0 0 > >> >> 0 0 0 0 0 > >> >> 2 0 0 1 0 > >> >> 3 0 0 0 0 > >> >> ; > >> > >> >> proc transpose data=have out=temp1; > >> >> run; > >> > >> >> data temp2; > >> >> set temp1; > >> >> if min(of col1-col4) + max(of col1-col4) ne 0; > >> >> run; > >> > >> >> proc transpose data=temp2 out=want (drop=_:); > >> >> run; > >> > >> >> HTH, > >> >> Art > >> >> --------- > >> >> On Fri, 9 Nov 2007 10:19:59 -0800, mhol...@CUESTA.EDU wrote: > >> > >> >> >Hello all, > >> >> > I am working with a dataset with ~200 variables and ~100 > >> >> >observations each. Most of the data are zeros. Is there a way to > >> >> >delete all of the variables that have all zero values? The If/then > >> >> >only seems to work for individual variables. Thanks. > >> > >> >> >Mike- Hide quoted text - > >> > >> - Show quoted text - > > > >a simpler form might be > > if range( of _numeric_ ); > > Wouldn't that fail to distinguish all =0 from all =1 etc.? >

Ahhh..... eagle eyed Howard Schrier, spotted my mistake (again!).

The correction is important but minor, and still avoids specifying column names > if range( of _numeric_ ); should have been if range( of _numeric_ ) and max( of _numeric_ ) ; Could also have been if min( of _numeric_ ) and max( of _numeric_ ) ;

Peter C