Question

Supposing I have the following rows:

VARIABLENAME   YYYYMM   XXXXYY   DATAOBS
variable_1   | 201101 | 111100 | 20
variable_2   | 201101 | 111100 | 20
variable_1   | 201102 | 111100 | 30
variable_2   | 201102 | 111100 | 30
variable_1   | 201101 | 222211 | 20
variable_2   | 201101 | 222211 | 20
variable_1   | 201102 | 222211 | 30
variable_2   | 201102 | 222211 | 30

I would like to know how to get the result:

VARIABLENAME      YYYYMM    XXXXYY   DATAOBS
variables_1and2 | 201101  | 111100 | 40
variables_1and2 | 201102  | 111100 | 60
variables_1and2 | 201101  | 222211 | 40
variables_1and2 | 201102  | 222211 | 60

where all the DATAOBS in rows containing variable_1 and variable_2 within each respective YYYYMM within in each respective XXXXYY are added and placed in a new row, variables_1and2, still within their respective YYYYMM and respective XXXXYY.

If it’s necessary to know, there are several hundred different YYYYMM and XXXXYY in the actual data I’m dealing with.

Thanks in advance!

Answer 1

It sounds like you just want

SELECT 'variables_1and2' variablename,
       yyyymm,
       xxxxyy,
       sum(dataobs) dataobs
  FROM your_table
 WHERE variablename in ('variable_1', 'variable_2' )
 GROUP BY yyyymm, xxxxyy

If the intention is to insert these rows back into your table, you’d need an INSERT

INSERT INTO your_table( variablename, yyyymm, xxxxyy, dataobs )
  SELECT 'variables_1and2' variablename,
         yyyymm,
         xxxxyy,
         sum(dataobs) dataobs
    FROM your_table
   WHERE variablename in ('variable_1', 'variable_2' )
   GROUP BY yyyymm, xxxxyy

Of course, it would be very unusual for it to make sense to store both detail and aggregate data in the same table. That would generally be a poor design.