Conditional merge If table exists

[Ronein]

Hello I have the following question.

I have a data set called 'tbl1' which contain fields: x,y1,y2.

Sometimes there another data set called "recovery_tbl' which contain fields: x,w,y1,y2

If in specific field all values are missing(null values in all columns) then If a recovery data set exists it will gather the values from there.

Data tbl1;
input ID  y1  y2;
cards;
1 10 .
2 20 .
3 30 .
4 40 .
5 50 .
;
Run;
Data recovery_tbl;
input ID w y1  y2;
cards;
1 5 10 100
2 6 20 200
3 7 30 300
4 8 40 400
5 9 50 500
;
Run;
PROC SQL;
	create table tbl1_new as
	select a.ID,a.Y1,coalesce(a.Y2,b.Y2) as Y2	   
	from tbl1 as a
	left join recovery_tbl as b
	on a.ID=b.ID
;
QUIT;

Can anyone how to perform this conditional merge. The merge need to be done only if 2 conditions happens: 1-There is a field that all values are null (missing values) 2- A recover table exists

[Astounding]

Two questions, just to clarify what is needed.

If only some of the y2 values are missing from TBL1, would you like to use TBL1 as is (including those missing values)?

Are you guaranteed that each table contains exactly one observation per ID?