How to select non-NA values for each ID if present otherwise select NA in Oracle?

Question

I have a df which can have single/several rows per ID. If it has multiple rows, only 1 row is populated with a non-NA value for a column. Sometimes all rows are NA. Using Oracle, I want to query out non-NA values if they exist otherwise the NA values as it is.

df:
ID Category  
1  NA  
1  A  
1  NA  
2  B  
2  NA  
3  NA  
4  C

I want to have the following as the final result df.

df:
ID  Category  
1  A  
2  B  
3  NA  
4  C

I tried joining with where clause non-NA but that is omiting ID 3.

Answer 1

Assuming there are not any nulls in the column category, for this sample data, you can use conditional aggregation:

select id, 
  coalesce(max(case when category <> 'NA' then category end), 'NA') category
from tablename
group by id

See the demo.