How to Fix Missing Values that Turned Numeric Variables to Character Variables

[unwashedhelimix]

I have a data set called "BASKETBALL_ST" and it looks something like this:

Age   Points   Assists   Rebounds   Steals

21          12.1        7.2               7.4                1.2

26          20.3        7.2               .                1.2

21          10.0          .               7.4                .

19          8.4        10.1               .               2.3

I want to run a proc means for it, but the variables with missing values (Assists, Rebounds, Steals) become character values, so they're not showing up on the proc means output.

How can I delete ONLY the cells with missing (not the whole row)? Age and points are numeric, so they are showing in the proc means report, as desired. However, I can't figure out how to delete the missing values without deleting an entire row. This is my code so far:

proc import datafile="C:\Users\John\Downloads\SASLearn\SASLearnOct\BASKETBALL_ST.xls"
   dbms=xls
   out=work.BASKETBALL_ST
   replace;

data work.cleanBASKETBALL_ST;
   set work.BASKETBALL_ST;
   if Assists=. then Assists=0;
   if Rebounds=. then Rebounds=0;
   if Rebounds=. then Rebounds=0;
run;

proc means data=work.cleanBASKETBALL_ST n mean median std min max range maxdec=2;
run;

This still yields a proc means report with only the variables Age and Points. At the end after deleting the missing values, Age & Points should have 4 observations, Assists should have 3, Rebounds should have 2, Steals should have 3.

After that is taken care of, I'm also stumped on how to utilize the input() statement to convert these back to numeric.

[Tom]

You cannot change the type of an existing variable.

So make new variables instead.

data cleanBASKETBALL_ST;
   set BASKETBALL_ST;
   _1 = input(assists,32.);
   _2 = input(rebounds,32.);
   _3 = input(steals),32.);
   rename _1=Assists _2=Rebounds _3=Steals 
     Assists=Assists_char Rebounds=Rebounds_char Steals=Steals_char
   ;
run;

Take a look at your XLS file and try to figure out why it has character values in some of the cells in the columns that you want to be numbers.  Where did it come from?  Can you get the data as a CSV file instead?  With a text file, such as a CSV file, you can write your own data step to read in the data lines.  That way you will have complete control over how the variables are defined.

[unwashedhelimix]

Thanks for the reply! I just saw it after revising my question a bit. I appreciate the guidance on the conversion, but I thought it would be a better report if the missing cells (not the entire row) were deleted.

[Tom]

---

@unwashedhelimix wrote:

Thanks for the reply! I just saw it after revising my question a bit. I appreciate the guidance on the conversion, but I thought it would be a better report if the missing cells (not the entire row) were deleted.

---

I do not know what you mean by that comment.   There is no need to delete observations (or values).  Can you explain what you meant more clearly?  

If you have data as you posted (with numeric variables) SAS will handle it properly. 

data have;
  input Age Points Assists Rebounds Steals;
cards;
21 12.1  7.2 7.4 1.2
26 20.3  7.2  .  1.2
21 10     .  7.4  .
19  8.4 10.1  .  2.3
;

For example the mean values will be calculated using only the non-missing values. 

The MEANS Procedure

Variable    N            Mean         Std Dev         Minimum         Maximum
-----------------------------------------------------------------------------
Age         4      21.7500000       2.9860788      19.0000000      26.0000000
Points      4      12.7000000       5.2883520       8.4000000      20.3000000
Assists     3       8.1666667       1.6743158       7.2000000      10.1000000
Rebounds    2       7.4000000               0       7.4000000       7.4000000
Steals      3       1.5666667       0.6350853       1.2000000       2.3000000
-----------------------------------------------------------------------------