Increase count for each run

[Babloo]

I have a situation where I need to increase the value by 1 to create a new variable(numeric) called 'count_execution' in one new dataset (want) if I execute any program from particular folder. Dataset 'want' is already created with 0 observation. So I want to insert value as 1 only for the first record in WANT dataset and for the subsequent records values should be dynamically inserted like 2, 3, 4, 5...

E.g.

Count_execution
1
2
3

[Kurt_Bremser]

data want;
set have end=eof;
output;
if eof
then do;
  count + 1;
  output;
end;
run;

[Babloo]

I ran the code below and I'm seeing the Output as 0 Observation  instead of 1 Observation.

Desired Output is 1 for the run_id if there is no Observation in the HAVE dataset. If any Observation then I want to add the run_id by 1.

data have;
input run_id;
datalines;
run;

data want;
set have nobs=nobservation nobs=n_obs;
if n_obs=0 then run_id=1;
else run_id+1;
run;

Log:

26         data have;
27         input run_id;
28         datalines;

NOTE: The data set WORK.HAVE has 0 observations and 1 variables.
NOTE: DATA statement used (Total process time):
      real time           0.00 seconds
      cpu time            0.00 seconds
      
29         run;

30         
31         data want;
32         set have nobs=nobservation nobs=n_obs;
33         if n_obs=0 then run_id=1;
34         else run_id+1;
35         run;

NOTE: There were 0 observations read from the data set WORK.HAVE.
NOTE: The data set WORK.WANT has 0 observations and 1 variables

I tried with your apporach like below as well and it is also not working.

26         data want;
27         set have end=eof;
28         output;
29         if eof
30         then do;
31           run_id+1;
32           output;
33         end;
34         run;

NOTE: There were 0 observations read from the data set WORK.HAVE.
NOTE: The data set WORK.WANT has 0 observations and 1 variables.

[Kurt_Bremser]

Yes, we need to take care for the empty dataset:

data want;
if _n_ = 1 and eof
then do;
  run_id = 1;
  output;
end;
set have end=eof;
output;
if eof
then do;
  run_id + 1;
  output;
end;
run;

[Babloo]

Thanks, your code works well for 0 observation in the HAVE dataset. In case if I want to grow the run_id by 1 for each run, then how can we tweak the code.?

For example, code below works well if I've Observation in HAVE and the code which you posted works well for 0 Observation. Now if I want to combine this behaviour how can we do it? I regret if I didn't meant this objective in my IP.

27         data have;
28         input run_id;
29         datalines;

NOTE: The data set WORK.HAVE has 1 observations and 1 variables.
NOTE: DATA statement used (Total process time):
      real time           0.00 seconds
      cpu time            0.00 seconds
      
31         ;

32         run;
33         
34         
35         
36         
37         proc sql noprint;
38         	select max(max(run_id)+1, 1) into :_newid
39         		from want;
40         	insert into want
41         		set run_id=&_newid;
NOTE: 1 row was inserted into WORK.WANT.

42         quit;
NOTE: PROCEDURE SQL used (Total process time):
      real time           0.00 seconds
      cpu time            0.00 seconds
      

43         
44         proc print data=want;
2                                                          The SAS System                     Friday, September 13, 2019 11:02:00 AM

45         run;

NOTE: There were 9 observations read from the data set WORK.WANT.

[FreelanceReinh]

Hi @Babloo,

Try this PROC SQL approach:

/* Assuming this already exists

data want;
stop;
run_id=.;
run;

*/

proc sql noprint;
select max(max(run_id)+1, 1) into :_newid
from want;

insert into want
set run_id=&_newid;
quit;

proc print data=want;
run;

If you run this code repeatedly, you'll see how dataset WANT is growing with RUN_ID=1, 2, 3, ...

[Babloo]

Thanks. Curious to know about data step method as well.

[FreelanceReinh]

Okay, then try this:

data want;
run_id=sum(n,1);
output;
stop;
set want nobs=n;
modify want;
run;

proc print data=want;
run;

Again, run the above code repeatedly (starting with the empty WANT dataset) to see the effect.

[Babloo]

Despite executing the following code (only the below code and any step prior to it) multiple times, I could see only the below NOTE. It's not adding any observation.

26         data want;
27         run_id=sum(n,1);
28         output;
29         stop;
30         set want nobs=n;
31         modify want;
32         run;

NOTE: The data set WORK.WANT has been updated.  There were 0 observations rewritten, 1 observations added and 0 observations 
      deleted.

[FreelanceReinh · Accepted solution]

---

@Babloo wrote:

Despite executing the following code (only the below code and any step prior to it) multiple times, I could see only the below NOTE. It's not adding any observation.

...
NOTE: The data set WORK.WANT has been updated.  There were 0 observations rewritten, 1 observations added and 0 observations 
      deleted.

---

Why do you think it's "not adding any observation" when the log says "1 observations added" and the PROC PRINT step (which you omitted from my code) confirms that?