在一个表中更新几百万行的列

Question

I need to update two columns from one table which conatins few million rows. What is the most efficient way for doing this?

It will be something like this:

update my_table set column_one = to_date('11/11/11', 'RR/MM/DD'),
                    column_two = to_date('11/11/11', 'RR/MM/DD');

or this:

update my_table set column_one = to_date('11/11/11', 'RR/MM/DD');
update my_table set column_two = to_date('11/11/11', 'RR/MM/DD');

Or there is a better way?

Answer 1

The most efficient way to update a table with million rows is not to perform update operation at all and rather create a new table with updated values.

1. Create a new table with updated columns while creation.
2. Notice all the Synonyms, Grants, Constraints, Indexes.
3. Drop the old table.
4. Rename new table to old table.
5. Recreated grants, synonyms, constraints, indexes.

Answer 2

If you don't want to create a new table, just break it up into chunks.

update my_table set column_one = to_date('11/11/11', 'RR/MM/DD'),
                    column_two = to_date('11/11/11', 'RR/MM/DD')
where rownum < 1000 
AND column_one <> to_date('11/11/11', 'RR/MM/DD')
AND column_two <> to_date('11/11/11', 'RR/MM/DD');

@Incognito's answer is probably more performant. Try this way if you're meaning for 'efficient' was less time spent on writing code.