compare value in same column

[jojozheng]

Hi, 

I have a dataset like below:

id   date   type

1    SEP    A

1    SEP    B

1    SEP    C

2    SEP   B

3   SEP     C

4   SEP    B

4   SEP    C

I want like below:  group by id,date, if this id in specific time, in all the type he has, if he has type A then his new type will be A and ignore other type.

so the important type should be A>B>C. 

id   date   type_new

1    SEP    A

2    SEP   B

3   SEP     C

4   SEP    B

how can I solve this? Thank you! 

[novinosrin]

if the type is actually A,B,C etc ranked i.e sorted by hierarchy as your sample is 

all you need is

data want;

set have;

by id date type;

if first.id;

run;

[jojozheng]

if it is not in that order, it shuffles like
1 sep b
1 sep c
1 sep a

what should I do?

[novinosrin · Accepted solution]

I would keep a look up table like, and it;s easy to update look up table when needed

Type  rank

a          1

b         2

c        3 

and look up using join, merge, hash,format etc

[jojozheng]

thank you!