how to use var_: in 'if then' statement

[nirsan]

Hi all, I have some variables named like this var_1, var_2, var_3.....var_50. Right now it is up to var_50, but anytime if a new case is added to the data, it can be var_51, var_52 etc. So, I use the term var_: when I run frequency or crosstabs to avoid repeated  codes and to avoid writing new codes whenever  a new case is added. But whenever I use 'var_:' in 'if then statement', it does not work. Any suggestion? One thing to mention that var_1 is a neumaric variable but I but similar character variable too (eg. hlt_1, hlt_2, hlt_3 etc.) with which I have the same problem. 

here is an example of code that I am using:

If var_: in (5) then flag=1;

if hlt_: in ('100897') then flag=1;

Thank you!

[s_lassen · Accepted solution]

You cannot use multiple variables on the left side of the IN operator, so you will have to change the expression, e.g. using WHICHN or WHICHC, e.g.

If whichn(5,of var_:) then flag=1;
if whichc('100897',of hlt_:) then flag=1;

[nirsan]

Hi, It worked. Thank you very much.

[Kurt_Bremser]

---

@nirsan wrote:
Hi, It worked. Can you please suggest what will be the code if I want to check the following condition?
If var_: not in (5) then flag=1;
if hlt_: not in ('100897') then flag=1;

Thank you!

---

That is so trivial that you can easily find the answer on your own. Just start the tool between your ears.

Hint: negate the whole condition.

[FreelanceReinh]

Hi @nirsan,

Alternatively, you can use the IN operator with an array:

data want;
set have;
array var_[50];
array hlt_[50];
if 5 in var_ then flag1=1;
if '100897' in hlt_ then flag2=1;
run;

To avoid the hardcoded array dimensions (50) you may want to replace the two ARRAY statements by

array var_[*] var_:;
array hlt_[*] hlt_:;

Note, however, that with this array definition (in general) the numeric suffix of a variable matches the corresponding array index only if the variables are ordered by the suffix in the dataset. Otherwise, for example, var_[3] could be var_2.

Edit:

---

@nirsan wrote:

(...) Right now it is up to var_50, but anytime if a new case is added to the data, it can be var_51, var_52 etc. (...)

---

That's why in many cases a "long" (vertical) data structure is preferable where a single variable, say, var holds the values of var_1, var_2, ... in different observations. Adding "a new case" then just means adding new observations (which is a routine operation), while the data structure remains unchanged.