Re: Backup and trimming data idea
Dear Rick,
Can say in short look for all the where and make sure there are all indexed? Is there any other indication when to put the order by to be part of the indexes? Priority always goes highest for '=' in the case of like?
This part is a bit new to me.
"So, if you can't figure out which table will be picked first, have two indexes so that the optimizer is free to pick:
INDEX(b, id, c) or (id, b, c), and
INDEX(b, c)
(Again, order of the fields is important.) ". If we look into all the SELECT, UPDATE, DELETE etc may not all involved the same number of keys. So for that case how to handle it.
Regarding the delete only two tables I can see for sure will have more then a million rows. So should I just partition this two tables alone and leave the rest as it is and do manual chunk deletes?