How to read from row 100 to row 200 from a large data?

[SeaMoon_168]

I would use the code to read row 100 to row 200 from a large dataset named "LargeData" below:

data subset;
set LargeData (firstobs=100 obs=200);
run;

Any other better options? Many thanks!

[mkeintz]

There are no better options, because the system will use the dataset pagesize to determine which page (think which physical block of disk storage) has obs 100.  And from there on it's sequential access through obs 200.     You would get almost as fast if firstobs=100100 obs=100200, because SAS doesn't read through the preceding disk blocks to get to the one of interest.

[SASKiwi]

In my experience FIRSTOBS and OBS are a very efficient way of pulling a small sample of rows from a large dataset so I don't see any need to find a 'better' way as long as it runs reliably and efficiently for your use case.

[Ksharp]

/*If your data engine allow randomly to access*/
data want;
do _n_=100 to 200;
 set sashelp.heart point=_n_;
 output;
end;
stop;
run; 

[mkeintz]

The "downside" of POINT= is that SAS has to recalculate the page containing the desired obs for each iteration of the DO loop, unlike sequential access.  This probably isn't too bad because that calculation is pretty quick, and the identified page for each iteration is likely to be cached from the prior iteration.