Processing Data to find best combinations

Input: Two Files: 1. Capacity - This contains (variable,limit) 2. [login to view URL] - This contains (id,variable,value) Output: [login to view URL] (id, variable, value) The process: 1. Import both Input Files 2. Create a subset of Score containing: a. The single highest value per variable b. The id must be unique 3. Write this to an output table 4. Remove the id selected above from the score table (so that it does not get used again). 5. If the count per variable in output table = [login to view URL] then remove the variable from the score table. (So that the variable is not selected above the capacity). 6. Cycle back to point 2 until there are no more records in score. 7. Write out the [login to view URL] file. Additional Note: Sample data: Capasity:- variable,limit AGT110,1875 AGT120,2187 AGT139,1562 AGT146,1875 AGT150,2187 AGT152,2187 AGT153,1875 AGT155,1875 AGT157,1250 Score:- clientId,variable,value 14313738,AGT160,0.06359 14317404,AGT162,0.07821 14362561,AGT155,0.21406 14321720,AGT161,0.32384 14314713,AGT162,0.07992 14369775,AGT163,0.1069 14353801,AGT153,0.06878 14365883,AGT165,0.08649125 14332650,AGT153,0.07421 The output file should only contain 31870 records. The first round should look as follows:- id,variable,value 14348744,AGT161,0.72528 14349046,AGT160,0.66255 14310070,AGT155,0.57827 14348800,AGT110,0.53863 14349598,AGT159,0.44097 14355688,AGT146,0.43844 14353679,AGT162,0.41105 14352158,AGT120,0.36759 14349991,AGT139,0.33552 14368664,AGT164,0.31323 14349627,AGT158,0.29891 14349800,AGT152,0.28776 14364479,AGT163,0.25394 14355732,AGT157,0.23732 14349641,AGT150,0.23238 14357995,AGT153,0.33374 14353811,AGT166,0.3119712 14349070,AGT165,0.3096063 Lastly the code must be fast very fast.