Now there is table A, which stores the core and busiest data of the entire business line. The approximate amount of data online is assumed to be 1 billion. Almost all concurrency is related to this table. Recent business needs , I want to add a field, but found that the impact is too great and cannot be changed directly. The compromise solution is to add a new table B to store the ID of A and the ones that need to be added. That field

Because I don’t have access to the online database, I can’t actually test it, so I don’t know what to use

SELECT * FROM A, B ON A.id = b.aid WHERE a.xx = 'zz' AND  b.xx = 'zz'

It is better to check a large range of B first, and then use the results in the search interval of B to check A. In this case, the first method may be very accurate, but I don’t know how much the performance will be worse.

The second range is not too precise