I have a data frame like this:
         MONTH  STORECODE   VALUE   GRP
   0     0      7           0       0
   1     0      7           44      0
   2     0      7           12      0
   3     0      8           26      0
   4     0      8           13      0
   5     0      8           0       0
   6     2      0           149     0
   7     2      0           14      0
   8     2      0           0       0
   9     2      1           76      0
  10     2      1           87      0
I want output data frame like this:
        MONTH  STORECODE   VALUE  GRP
 0      0      7           56     0
 1      0      8           39     0
 2      2      0           163    0
 3      2      1           163    0
I.e. merge rows into a single on adding VALUE's value if the other 3 values are the same. 
For the first row in output, as the first 3 rows in the input df have the same values other than the VALUE's value, these 3 merged into a single row and VALUE's value updated on taking the sum of corresponding values (0+44+12=56).