Avoid numeric conversion warning when using format INVALUE to recode

[SAShole]

Hi SAS frens,

I'd like to use PROC FORMAT to recode Weight into an Injection Volume category but I get a NOTE (considered an Error by my company and will not pass validation) on the Numeric to Character conversion. How do I avoid this? I'd like to not use IF/THEN logic because my list is quite long.

data have;
 input id weight @;
 datalines;
1008 31.2
1009 47.8
1010 57.5
1011 95.5
1012 106.1
1013 110.0
1014 125.8
1015 166.3
 ;
run;

proc format;
 invalue vol3_ 
	 30.0 - 40.8  = 0.2
	 40.9 - 57.4  = 0.3
	 57.5 - 74.1  = 0.4
	 74.2 - 90.8  = 0.5
	 90.9 - 107.4 = 0.6
	107.5 - 124.1 = 0.7
	124.2 - 140.8 = 0.8
	140.9 - 157.4 = 0.9
	157.5 - 174.1 = 1.0
	174.2 - 180.0 = 1.1;
run;

data want;
	set have;
	inj_vol=input(weight, vol3_.);
run;

Thank you 😁

[ballardw · Accepted solution]

INPUT function expect text values so control the conversion to text of your existing numeric value with the PUT function and an appropriate format.

The -L option with the Put function left justifies the value so there are not leading spaces which might confuse your informat.

data want;
	set have;
	inj_vol=input(put(weight,best8. -L), vol3_.);
run;

OR create a Format and INPUT the formatted value.

inj_vol =  input (put(weight,yourvolfmt. -L),8.);

You want to be very cautious about ranges of character values, which the Informat actually stores for the start and end values. Note that your informat will treat a value like 40.8999999 as if it is 40.8 because of the rules for comparing character values.

You may want to specify an OTHER option in the value ranges with the _error_ assignment to create notes in the log about unexpected values.

Example, note the added row of data in the Have with the value of 3000.

data have;
 input id weight @;
 datalines;
1008 31.2
1009 47.8
1010 57.5
1011 95.5
1012 106.1
1013 110.0
1014 125.8
1015 166.3
1016 3000
 ;
run;

proc format library=work cntlout=work.cntlout;
 invalue vol3_ 
	 30.0 - 40.8  = 0.2
	 40.9 - 57.4  = 0.3
	 57.5 - 74.1  = 0.4
	 74.2 - 90.8  = 0.5
	 90.9 - 107.4 = 0.6
	107.5 - 124.1 = 0.7
	124.2 - 140.8 = 0.8
	140.9 - 157.4 = 0.9
	157.5 - 174.1 = 1.0
	174.2 - 180.0 = 1.1
   other = _error_
   ;
run;

data want;
	set have;
	inj_vol=input(put(weight,best8. -L), vol3_.);
run;