SQL Inner Join with and without "Inner Join" phrase?

[Mirisage]

Hi SAS Forum,

I took the attahced two data sets from "SUGI 27 Paper 70-27 An Introduction to SAS PROC SQL by

Harrington, T. J. (Acknowledged thankfully).

• I first ran the proc SQL inner join with "inner join" phrase.
• Then ran without "inner join" phrase.
• Then compared the two output files using "proc compare".
• Results show both produced identical results.

Question:

For the given small two data sets, both appraoches produced identical results with "inner join" phrase or without it.

Can I however generalize this conclusion for other situaitions invovling large data sets?

Thanks

Mirisage

[Murray_Court · Accepted solution]

My understanding was that the  difference between using the inner join phrase and other methodology what that of efficiency, not output. I think it is that the inner join phrase uses a more elegant method than just a "where a.id=b.id" subsetting clause, as it does not simply construct a cartesian product and weed out all of the irrelevant matches.

[PGStats]

The constructs

select ... from A, B where A.var=B.var

and

select ... from A inner join B on A.var=B.var

are perfectly equivalent. However, in the second case you can replace the word inner by left or right, and by cross or natural if you also remove the on clause.

PG

[Mirisage]

Hi Murray_Court and PGStats,

Many thanks to both of you for this help.

Regards

Mirisage