Left outer join syntax without using proc sql

[GN0001]

Hello team,

How can we do Left outer join in SAS without using sql?

Can someone help me with this?

Regards,

blueblue

[jimbarbour · Accepted solution]

Sure.  Just do what we did before there was Proc SQL; use a Merge.  

Here, I'm going to do essentially a left join.  I'm joining in phone_number, but I want to keep the individual even if I can't get a phone number for them. Regrettably, I couldn't get a phone number for Joe Biden and Jeff Bezos for this example.  🙂

DATA	Have;
	LENGTH	ID $4	Name $64	Address $64	City $64;
	INFILE	DATALINES	MISSOVER	DSD	DLM='09'x;
	INPUT	ID $	Name $		Address $	City $;
DATALINES;
2418	Barbara Ann	1515 Android Lane	Milwaulkee
2724	Walt Disney	48 IBM Way	Anaheim
5429	John Q Public	24 Apple Str	Des Moines
9527	Joe Biden	1600 Pennsylvania Avenue NW	Washington D.C.
9610	Jeff Bezos	1 Blue Origin St	Kent
;
RUN;

DATA	Have2;
	LENGTH	ID $4	Phone_Number $12;
	INFILE	DATALINES	MISSOVER	DSD	DLM='09'x;
	INPUT	ID $	Phone_Number $;
DATALINES;
2418	555-222-3333
2724	888-444-111
5429	999-876-5309
;
RUN;

PROC	SORT	DATA=Have;
			BY	ID;
RUN;

PROC	SORT	DATA=Have2;
			BY	ID;
RUN;

DATA	Want;
	MERGE	Have	(IN=Have)
			Have2	(IN=Have2);
			BY	ID;
RUN;

Results below.  Note how individuals without a phone number are still present in the results.

Jim

[GN0001]

This is a great input.

Thanks for this response:

Does this 

	MERGE	Have	(IN=Have)
			Have2	(IN=Have2);

Mean all the rows in the output, regardless whether there is a match or not.

What is the syntax if you want to bring all the cases?

Regards,

blueblue

[Kurt_Bremser]

Please do study the documentation of the MERGE statement.

In short, MERGE reads observations side-by-side. HOW it does that is controlled by an (optional) BY statement; subsequent conditions can be used to keep observations selectively.

[GN0001]

Hello,

I have studied the documentation. It was explained not very easy to grasp. I found a white paper in SAS community which helped me to understand the topic. Aslo, there is a forum member named Jimbourbar explained it very well.

so I have the concept now and tomorrow, I have to test the syntax.

Thanks for all.

blueblue

[jimbarbour]

---

@GN0001 wrote:

Does this 

	MERGE	Have	(IN=Have)
			Have2	(IN=Have2);

Mean all the rows in the output, regardless whether there is a match or not.

Yes.  Since I'm not using any conditions to limit the records in the output Data set, all rows are kept. 

Jim