simple array question

[Angmar]

New SAS user - trying to make an array that creates a new variable with observations as '1' when it finds either 9359 or 9351 (character, not numeric) in the array variables.

Here's my code

data cuthip;
set allrecords;
array procedure (3) $ prcode1-prcode3;
do x=(1-3);
if procedure(x) = ('9351'OR '9359') then procedure (x)=1;
end;
run;

I keep getting the error statement:

NOTE: Character values have been converted to numeric values at the places given by: (Line):(Column).
336:4 336:20 336:29
NOTE: Numeric values have been converted to character values at the places given by: (Line):(Column).
336:42
ERROR: Array subscript out of range at line 336 column 4.
ikn=1 prcode1=1257 prcode2=1165 prcode3=1165 rescode=angmar cats=amino x=-2 _ERROR_=1 _N_=1
NOTE: The SAS System stopped processing this step because of errors.
NOTE: There were 1 observations read from the data set WORK.ALLRECORDS.
WARNING: The data set WORK.CUTHIP may be incomplete. When this step was stopped there were 0 observations
and 7 variables.
WARNING: Data set WORK.CUTHIP was not replaced because this step was stopped.
NOTE: DATA statement used (Total process time):
real time 0.02 seconds
cpu time 0.01 seconds
 

[KachiM]

Try assigning a character constant to procedure[x] in:

if procedure(x) = ('9351'OR '9359') then procedure (x)=1;

with '1'.

[Astounding · Accepted solution]

As a new user, your errors are understandable.  Here are a few issues with your code.

do x=(1-3);

1-3 is -2.  You could just as easily have coded:

do x=-2;

To get x to go from 1 to 3:

do x=1 to 3;

Next, to compare to a list of values:

if procedure(x) in ('9351' '9359') then ...............

Finally, you said you wanted to create a new variable.  But instead, your code re-assigns a value to an existing variable.  This would be an improvement:

if procedure(x) in ('9351' '9359') then newvar=1;

[Angmar]

Thanks - little tips like that really help me out in understanding SAS language.

What if I wanted to keep only those records/observations (the whole row of data) with '9359' or '9351'? Would I use a keep statement?

[Astounding]

To keep or remove observations (based on a calculated variable) use an IF statement.  IF without THEN is totally different than IF-THEN.  To subset the observations, you could add this as the final statement in the DATA step:

if newvar=1;

Often an IF statement appears earlier in the DATA step.  It is usually most efficient to place it early.  But here you need to calculate NEWVAR first.