I am sorry if I am asking a stupid question, but my developers keep bugging me...

we are doing data updates in tables. and sometimes it’s 2 records, sometimes - 5000, and sometimes - 900,000. is there any kind of “best practice” at what number of records we should go from using option 1 to option 2 and than to option 3:
1. update set (few records at the same time);
2. import with insert_update (most or all records will be updated);
3. export, update values in export file, load replace;

Thanks in advance

Option 3 sounds like an oldfashioned mainframe solution to me. This involves a lot of hassle to be sure that no other process is using that table. When more than 80% of the data has to be changed you could consider it, but if you can find the (processing) time to do the job without the hassle, do it the normal SQL way to be sure and "rollforwardable"..... (you like the new word I just invented?)