Application of Filters to Multiway Joins in MapReduce

Algorithm 1

Finding target reducers.

⊳𝑛: the number of join attributes

 ⊳𝑗𝑜𝑖𝑛𝐴𝑡𝑡𝑟: an array of join attribute values (some values may be missing)

 ⊳𝑟𝑒𝑝𝑙: an array of replication factors for each join attribute

(1) procedure  FindTargetReducers(𝑗𝑜𝑖𝑛𝐴𝑡𝑡𝑟 [1,…,𝑛],  𝑟𝑒𝑝𝑙[1,…,𝑛])

(2)    𝑟𝑒𝑑𝑢𝑐𝑒𝑟𝐿𝑖𝑠𝑡←∅

(3)    𝑐𝑜𝑜𝑟𝑑←∅

(4)    for  𝑖=1 to 𝑛  do

(5)     if  𝑗𝑜𝑖𝑛𝐴𝑡𝑡𝑟[𝑖] is not null  then

 ⊳𝐺𝑒𝑡𝑃𝑎𝑟𝑡𝑖𝑡𝑖𝑜𝑛(): returns a number in the range [0 and (repl [𝑖]−1)]

(6)     𝑐𝑜𝑜𝑟𝑑[𝑖]←𝐺𝑒𝑡𝑃𝑎𝑟𝑡𝑖𝑡𝑖𝑜𝑛(𝑗𝑜𝑖𝑛𝐴𝑡𝑡𝑟[𝑖],𝑟𝑒𝑝𝑙[𝑖])

(7)     else

(8)     𝑐𝑜𝑜𝑟𝑑[𝑖]←0

(9)     end if

(10)    end for

(11)   while  𝑡𝑟𝑢𝑒  do

(12)     𝑟𝑖𝑑←𝐶𝑜𝑜𝑟𝑑𝑇𝑜𝑅𝑒𝑑𝑢𝑐𝑒𝑟(𝑐𝑜𝑜𝑟𝑑,𝑟𝑒𝑝𝑙)

(13)     𝑟𝑒𝑑𝑢𝑐𝑒𝑟𝐿𝑖𝑠𝑡←𝑟𝑒𝑑𝑢𝑐𝑒𝑟𝐿𝑖𝑠𝑡 ∪ 𝑟𝑖𝑑

(14)    if  𝐼𝑛𝑐𝑟𝐶𝑜𝑜𝑟𝑑(𝑐𝑜𝑜𝑟𝑑,𝑗𝑜𝑖𝑛𝐴𝑡𝑡𝑟,𝑟𝑒𝑝𝑙)=𝑓𝑎𝑙𝑠𝑒  then

(15)     break

(16)    end if

(17)   end while

(18)    return  𝑟𝑒𝑑𝑢𝑐𝑒𝑟𝐿𝑖𝑠𝑡

(19) end procedure