Uninitialized Variables Troubleshooting

[EmmaW]

I've been trying to perform a row sum for 14 variables in my data set. 

Here's my code: 

DATA temp;
set work.silver2;
sum_chronic=q2a_temp + q2b_temp + q2c_temp + q2d_temp + q2e_temp + q2f_temp + q2g_temp + q2h_temp + q2i_temp + q2j_temp + q2k_temp + q2l_temp + q2m_temp + q2n_temp;
run;

Every time I run the data, it tells me that the variables are unintialized, but this is the same "data" line and "set" that I collapsed the variables into earlier.  Any advice for how to fix this? I'm very confused.  

[SASKiwi]

Please post your complete SAS log including all messages so we can see more clearly what is going on.

[EmmaW]

Here's the log output (I used the sum function instead, as another person recommended in the comments): 

2805 DATA temp;
2806 set work.silver2;
2807 chronic_disease=.;
2808
2809 IF q2a_temp = 1 THEN _q2a_temp = 1;
2810 IF q2b_temp = 1 THEN _q2b_temp = 1;
2811 IF q2c_temp = 1 THEN _q2c_temp = 1;
2812 IF q2d_temp = 1 THEN _q2d_temp = 1;
2813 IF q2e_temp = 1 THEN _q2e_temp = 1;
2814 IF q2f_temp = 1 THEN _q2f_temp = 1;
2815 IF q2g_temp = 1 THEN _q2g_temp = 1;
2816 IF q2h_temp = 1 THEN _q2h_temp = 1;
2817 IF q2i_temp = 1 THEN _q2i_temp = 1;
2818 IF q2j_temp = 1 THEN _q2j_temp = 1;
2819 IF q2k_temp = 1 THEN _q2k_temp = 1;
2820 IF q2l_temp = 1 THEN _q2l_temp = 1;
2821 IF q2m_temp = 1 THEN _q2m_temp = 1;
2822 IF q2n_temp = 1 THEN _q2n_temp = 1;
2823
2824 SUMRACE=
2824! sum(_q2a_temp,_q2b_temp,_q2c_temp,_q2d_temp,_q2e_temp,_q2f_temp,_q2g_temp,_q2h_temp,_q2i_tem
2824! p,_q2j_temp,_q2k_temp,_q2l_temp,_q2m_temp,_q2n_temp);
2825 RUN;

NOTE: Variable q2a_temp is uninitialized.
NOTE: Variable q2b_temp is uninitialized.
NOTE: Variable q2c_temp is uninitialized.
NOTE: Variable q2d_temp is uninitialized.
NOTE: Variable q2e_temp is uninitialized.
NOTE: Variable q2f_temp is uninitialized.
NOTE: Variable q2g_temp is uninitialized.
NOTE: Variable q2h_temp is uninitialized.
NOTE: Variable q2i_temp is uninitialized.
NOTE: Variable q2j_temp is uninitialized.
NOTE: Variable q2k_temp is uninitialized.
NOTE: Variable q2l_temp is uninitialized.
NOTE: Variable q2m_temp is uninitialized.
NOTE: Variable q2n_temp is uninitialized.
NOTE: Missing values were generated as a result of performing an operation on missing values.
Each place is given by: (Number of times) at (Line):(Column).
596 at 2824:16
NOTE: There were 596 observations read from the data set WORK.SILVER2.
NOTE: The data set WORK.TEMP has 596 observations and 327 variables.
NOTE: DATA statement used (Total process time):
real time 0.04 second

When I was cleaning up the variables before trying to do this sum, I was using the format: 

2702
2703 DATA temp;
2704 Set work.silver2;
2705 q2m_temp=.;
2706 IF (Q2m=1) then q2m_temp=0;
2707 IF (Q2m=2) then q2m_temp=1;
2708 IF (Q2m=98) then q2m_temp=98;
2709 IF (Q2m=99) then q2m_temp=99;
2710 IF (Q2m=901) then q2m_temp=901;
2711 LABEL q2m_temp="q2m_temp";
2712 RUN;

That seemed to be working just fine and I was getting the output that I needed using the same recode format for all 14 variables.  The data set doesn't have missing variables, although once I get a sum formula that works, I will have to initialize the 98, 99 and 901 to missing before adding (I think).  I'm very new at this, so any advice is welcome

[Tom]

If it did not generate the uninitialized variable message before then something has changed.

If the code did not change then the dataset SILVER2 has changed. It no longer has those variables.

[ballardw]

Show us the result of Proc Contents on the data set Work.silver2. Just need the section of the variable names and types.

Hint: Unless your desired behavior is to have a missing value when any of the variables on that addition statement are missing I would suggest using the SUM function which will total the non-missing values. The + operator will return missing if ANY of the values is missing. Note that this means that if every observation has one or more missing values for the q2 _temp variables then Sum_chronic would be missing for every observation.

I don't know who suggested naming variables q2a_temp q2b_temp etc. If those had been named Temp_q2a Temp_q2b etc. You might be able to use the list short cut of temp_q2:  to use all the variables whose names start with temp_q2 so that long statement could reduce to

sum_chronic = sum(of temp_q2:);

And the A, B, C etc suffixes might be better, at least in terms of SAS coding as 1, 2, 3 such as Temp_q2_1 Temp_q2_2 Temp_q2_3 because there are places where enumerated lists such at Temp_q2_3 - Temp_q2_10 to use the 3rd through 10th of the temp q2 variables might be handy. The character suffix doesn't play nice with sequences in names.