Question

I have a queries like below. Can these be optimized or can someone suggest a how to index?

SELECT  t1.`Geo ID`, 
        t1.`Sub Unit Geo ID`, 
        t1.`Sub Unit SW ID`, 
        t1.`Building No`, 
        t1.`Building Name`, 
        t1.Road, 
        t2.ID `Matching NEW_ID`, 
        t2.Sub, 
        t2.SUB_BUILDING_NAME, 
        t2.BUILDING_NAME,
        t2.BUILDING_NUMBER, 
        t2.THOROUGHFARE, 
        t2.E - t1.Easting `East Difference`, 
        t2.N - t1.Northing `North Difference` 
FROM    upcdata t1 JOIN 
        newer t2    ON  (t2.E * 1000) BETWEEN t1.Easting - 25000 AND t1.Easting + 25000 
                    AND (t2.N * 1000) BETWEEN t1.Northing - 25000 AND t1.Northing + 25000 
                    AND t1.Road = t2.THOROUGHFARE 
                    AND t1.`Building Name` = t2.BUILDING_NAME 
                    AND t1.`Building Name` <> ""
ORDER BY    t1.`Geo ID`

Answer 1

Indexes aren’t used for calculated values, but you can alter your query so solve that!
Change your between to put the multiplication on the values:

....
join newer t2
    ON t2.E BETWEEN (t1.Easting - 25000)/1000 AND (t1.Easting + 25000)/1000
    AND t2.N BETWEEN (t1.Northing - 25000)/1000 AND (t1.Northing + 25000)/1000
....

The comparison is mathematically identical, but it means the between has 2 fixed values to look up on, rather than having to multiply the t2 values for every row and not use an index because of it.

Even without indexes, just avoiding millions of computations will make it much faster