SQL decode column value to another column

I am using PL/SQL 12.
I want to decode a column value to match another column value. I take the highest value of the the columns SP_RANK and MOODY_RANK and call it RANK. I want this column to instead of values have the corresponding value in either SP or MOODYS depending on which value is highest.

Below my code and a snippet of the data.

SELECT rv.updat
  , rv.instrumentid
  , greatest(nvl(rv.sp_rank, '-1')
  , nvl(rv.moody_rank, '-1')) AS "Rank"
  , rv.sp
  , rv.moodys
  , rv.SP_Rank
  , rv.Moody_Rank
FROM ci.ratings_view rv
ORDER BY rv.updat ASC

Table:

INSTRUMENTID | UPDAT      | Rank | SP  | MOODYS | SP_RANK | MOODY_RANK 
-------------+------------+------+-----+--------+---------+-----------
0203041      | 26/09/2019 | 100  | AAA | Aaa    | 100     | 100       
0203378      | 26/09/2019 | 100  | AAA | Aaa    | 100     | 100        
0203734      | 26/09/2019 | 100  | AAA | Aaa    | 100     | 100       
0204196      | 26/09/2019 | 100  | AAA | Aaa    | 100     | 100       
0204277      | 26/09/2019 | 100  | AAA | Aaa    | 100     | 100       
0413194      | 26/09/2019 | 75   | A   | NR     | 75      | -1        
0413216      | 26/09/2019 | 75   | A   | NR     | 75      | -1        
0413240      | 26/09/2019 | 75   | A   | NR     | 75      | -1        
0460583      | 26/09/2019 | 100  | AAA | NR     | 100     | -1        
0460656      | 26/09/2019 | 100  | AAA | NR     | 100     | -1        
0471534A     | 26/09/2019 | 100  | AAA | WR     | 100     | -1        
0491438      | 26/09/2019 | -1   | NR  | NR     | -1      | -1        

So instead of 100 it should say AAA and instead of 75 A, etc.

Answer

You can do this with a CASE statement:

select 
  rv.updat, 
  rv.instrumentid, 
  greatest(nvl(rv.sp_rank,'-1'), 
  nvl(rv.moody_rank,'-1')) as "Rank",
  rv.sp, 
  rv.moodys, 
  rv.SP_Rank, 
  rv.Moody_Rank, 
  CASE 
    WHEN COALESCE(rv.moody_rank, -1) > COALESCE(rv.SP_Rank, -1) THEN rv.moodys 
    ELSE rv.SP 
  END AS NewRank
from ci.ratings_view rv
order by rv.updat asc

You may need to add some extra logic to handle ties — i.e. which rank to use (SP or Moodys)