Question

What is the best practice to update a table record most effectively (in my case with a primary key), when not all values are present?

Imagine:

PRIMARY_KEY1, COLUMN_2, COLUMN_3, COLUMN_4, COLUMN_5, COLUMN_6, ...

I always get tuples like (PRIMARY_KEY1, COLUMN_5, COLUMN_4) or (PRIMARY_KEY1, COLUMN_2, COLUMN_6, COLUMN_3) and want to just update them in the fastest way possible without having a database lookup for all other values.

Since I have to to this very fast, I would like to use something like batches for prepared statements in order to prevent massive database requests.

Thanks for all replies!

Answer 1

You can ‘cheat’ by expecting SQL to fill in the values at row-access time. Eg, this type of statement:

UPDATE MyTable SET (column_1, column_2, ..., column_6)
                      = (COLAESCE(@suppliedValue1, column_1),
                         COLAESCE(@suppliedValue2, column_2),
                         ...,
                         COLAESCE(@suppliedValue6, column_6))
WHERE primary_Key1 = @primaryKey

Then, when filling out the parameters, just leave anything unsupplied null… and you should be good.