What would be an efficient way of preventing sequential scan if data is partitioned by list, when it comes to updating multiple partitions with a condition that is not referring to the partition key?
My table contains
- the partition key is
IndividualNameis a search key.
SomeComputedValueis the value that will be updated across different groups.
SomeTypeis an additional search parameter
I know Searching via
GroupName would be faster, but what if my data end up requiring to search for
Is there anything wrong if I end up requiring to update
SomeComputedValue across 10 to 100 partitions, with 1 to 10 million rows each?
Should an index be used or avoided?
My search would be something like
SomeType in the update statement. But after that most of the time it will be a
GroupName search for select and no more update.
I feel like the index would be more detrimental.