Normalizing Data on a Table

Question

I need help to transform data on a table, reducing a series of columns to one single column. An example follows below:

Frequency_1 integer,
Frequency_2 integer,
Frequency_3 integer,
Frequency_4 integer,

These columns currently hold 1 or 0. Only one column will hold 1.

The new column should be defined as

Frequency integer

And this new column should hold a value between 1 and 4, depending on which of the old columns had its value = 1.

Could you suggest an SQL command to accomplish this?

Answer 1

You could come up with something more complicated if you want, but why not just do this?

SELECT Frequency_1 +
       (Frequency_2 * 2) +
       (Frequency_3 * 3) +
       (Frequency_4 * 4) AS Frequency

To actually make the change, you can create the column first, update the value in the new column, then delete the old columns.

Answer 2

SELECT
    CASE WHEN Frequency_1 = 1 THEN 1 
         WHEN Frequency_2 = 1 THEN 2 
         WHEN Frequency_3 = 1 THEN 3 
         WHEN Frequency_4 = 1 THEN 4 
         ELSE 0 END AS Frequency
FROM TABLE

Answer 3

update table_name
  set frequency = 
    case when frequency_1 = 1 then 1 else
      case when frequency_2 = 1 then 2 else
        case when frequency_3 = 1 then 3 else
          case when frequency_4 = 1 then 4 
          end
        end
      end
    end

Answer 4

Like this:

select Frequency =
  Frequency_1 * 1 +
  Frequency_2 * 2 +
  Frequency_3 * 3 +
  Frequency_4 * 4
from ATable

or this:

select Frequency = case
  when Frequency_1 = 1 then 1
  when Frequency_2 = 1 then 2
  when Frequency_3 = 1 then 3
  when Frequency_4 = 1 then 4
  else 0
end
from ATable