Assume table T with columns: i binary, j int, k int, x varchar(5), y varchar(10).
Following are the indexes that this table have on it:
Idx1 on columns i,j,x
Idx2 on columns i,k,x
Idx3 on columns i,j,y
Idx4 on columns i,k,y
Note that the leading column in all the indexes is 'i'. There are different queries whose where clauses include conditions on columns (i,j,x), (i,k,x), (i,j,y) and (i,k,y). These queries select around 10-20% of the total no. of rows.
As we are maintaining 4 indexes, DML operations are taking longer time to process. If we drop all these indexes and create another one, say Idx on columns i,j,k,x,y - does this going to have any(too much) -ve effect on my SELECT queries?