How to sum observation according to multiple conditions

[yanshuai]

Hello,

I have a table such as

ID           Sales     Year     Sales_classification

1             12         2008              4567

1              11        2008               7890

1              13        2008               5678

1              14        2009               4567

1              15        2009               7890

1               16       2009               5678

2                8       2008               4567

2              1        2008               7890

2              3        2008               5678

2              4        2009               4567

2              5        2009               7890

2               6       2009               5678

Condition includes:

IF

With the same ID

With the same YEAR

Sales_classification first digit larger than 6

THEN

sum the SALES meet all the conditions above

This may be done with or without CONDITION proc, there may be several ways and logic to do this.

Would you please help me with it?

Thank you very much!

[PGStats]

You could do:

proc sql;
create table want as
select ID, year, sum(Sales) as sumSales
from have
where first(cats(Sales_classification)) > "6"
group by id, year;
quit;

[yanshuai]

OMG

You are awesome! It exactly does what I want. It is so neat and clean.

But the outcome table only shows the observation meet the condition. What if I also want those obs show "0" if do not meet the condition?

[PGStats · Accepted solution]

Then switch to:

proc sql;
create table want as
select 
    ID, 
    year, 
    sum(case when first(cats(Sales_classification)) > "6" 
        then Sales 
        else 0 end) as sumSales
from have
group by id, year;
quit;

[yanshuai]

Great!

Thank you!