what is the point of a one on one reading in data step when it gives up wrong combined data set?

[Nietzsche]

Hi I am reading page 181 of the official specialist prep guide on Using One-to-One Reading to Combine Data Sets

Basic I have two data sets, one is 11 obs and the other is 9 obs and I combine them with a simple data step.

data work.one2one;
set cert.patients;
set cert.measure;
run;

the result is the one on the right hand side.

As you can see, the ID column in the second data set overrides the ID column in the first data set in the PDV, giving an output with wrong combined observation data.

So what is the point of one on one reading in data manipulation?

I have attached the two data sets.

[Tom · Accepted solution]

I don't think that they really care about what is in the data, they are just using the two datasets to illustrate the point that a data step will stop when it reads past the end of an input.

Most SAS data steps do not end at the last line of the code in the data step.  Instead they end when they read past the end of the input date (either an input dataset in a SET/MERGE/UPDATE statement or an input file in an INPUT statement).

Try it yourself to see:

data test;
  put 'BEFORE the SET statement ' _n_=  age= ;
  set sashelp.class ;
  put 'AFTER the SET statement ' _n_= age=;
run;

But the fact that the two datasets have variables in common is another important thing to learn.  The value that is available in the data step is the last value read in.  So if you change the order of the two SET statements the output will have the same number of observations, but the values in each observation might be different because the data was read in a different order.