topic Re: Data Manipulation in SAS Programming

Data Manipulation

Ujjawal — Thu, 09 Jun 2016 18:01:00 GMT

I want to add serial number to the unique values of each group. See the snapshot below-

data xyz;
input x $;
cards;
AA
AA
AA
BB
BB
;
run;

proc sql;
select a.x, b.N from xyz a
inner join
(select x, monotonic() as N
from (
select distinct x
from xyz)) b
on a.x=b.x;
quit;

How we can do it with data step?

Re: Data Manipulation

Reeza — Thu, 09 Jun 2016 18:20:18 GMT

http://www.ats.ucla.edu/stat/sas/faq/enumerate.htm

Use BY group processing, see a detailed write up in the link above.

Re: Data Manipulation

Ujjawal — Thu, 09 Jun 2016 18:26:03 GMT

I have tried but didn't get any success.
data temp;
set xyz;
if first.x then z=1;
else z+1;
by x;
proc print;
run;

It increments the value. It's not my desired output.

Re: Data Manipulation

Reeza — Thu, 09 Jun 2016 18:38:09 GMT

Close but no cigar:

data temp;
set xyz;
by x;
retain z 0;
if first.x then z+1;
run;

proc print;
run;

Re: Data Manipulation

Ujjawal — Thu, 09 Jun 2016 19:01:28 GMT

Thanks a ton. It works like a charm. How it works? I failed to understand how it increments with "if first.x then z+1;" and how it keeps the same value to values of a same group?

Re: Data Manipulation

ballardw — Thu, 09 Jun 2016 19:06:50 GMT

RETAIN is a datastep instruction that says to keep the value of a variable from record to another. This will not typically work if the variable(s) named in the Retain statement exist in a data set on the SET statement as they would be read from the set.

Using a statement like:

Z+1; /* note there is no = in the assignment*/

implies that Z should be retained.

Re: Data Manipulation

Ujjawal — Thu, 09 Jun 2016 19:14:06 GMT

Thanks for your explanation. I am confused why ELSE statement was not required. See my code below-

data example22;
set xyz;
if first.x then N+1;
else N=N;
by x;
proc print;
run;

Re: Data Manipulation

Reeza — Thu, 09 Jun 2016 20:05:34 GMT

BY group processing:

http://support.sas.com/documentation/cdl/en/lrcon/62955/HTML/default/viewer.htm#a001283274.htm

RETAIN

keeps the value from row to row

Re: Data Manipulation

ballardw — Thu, 09 Jun 2016 20:45:03 GMT

Because you wanted the same value of the counter for all records with the same value of x you don't want an Else. The retain keeps the value once assigned.

The ELSE clause would be for something you only want to do for records that are not the first.x.

See this thread for one example of Else with First.var processing to do something just a little more complicated that your problem:

https://communities.sas.com/t5/SAS-Data-Management/Creating-a-counter-for-the-same-hospitalization/m-p/276010#M7767