topic Re: Doubt in New SAS User

Doubt

Pooja98 — Fri, 16 Oct 2020 17:21:55 GMT

which statement is efficient between where and if statement??

Re: Doubt

ballardw — Fri, 16 Oct 2020 17:33:38 GMT

Define "is efficient".

IF statements can check conditions that WHERE can't. Variables used in a Where expression have to exist in a source data set.

WHERE can use some operations that IF can't. Where is usable in most procedure but IF isn't.

Re: Doubt

mkeintz — Fri, 16 Oct 2020 20:52:59 GMT

In addition to @ballardw 's comment, there is another important difference between WHERE and IF - namely that where will outsource the filtering to the data engine, while if only filters once the data observation has been passed to the data step. This data engine might be the sas engine for various types of sas datasets, or a database for which sas/access has an engine (see SAS/ACCESS Software Access Engines).

The fact that where-filtering is outsourced is not only why you can use where in your proc freq's, proc reg's, proc sort's, etc., but it also can make the where statement (compared to a subsetting if) a more efficient filter in a data step.

Re: Doubt

Ksharp — Sat, 17 Oct 2020 09:45:37 GMT

Why not make a dummy data to test them ?
WHERE could get faster if it is small data.
IF could get faster if it is big data.

Re: Doubt

andreas_lds — Mon, 19 Oct 2020 09:57:17 GMT

@Pooja98 wrote:
which statement is efficient between where and if statement??

You have to use where instead of subsetting-if, whenever you are using by-group-processing or end=somevar in the set statement. Execute the following steps and have a look at the log:

data girls1;
   set sashelp.class end=lastObs;
   where Sex = 'F';
   if lastObs then put 'Just read ' _n_ 'obs.';
run;

data girls2;
   set sashelp.class end=lastObs;
   if Sex = 'F';
   if lastObs then put 'Just read ' _n_ 'obs.';
run;