There are a couple of factors to consider:
1) Is this a deployed app? If not, it runs in a test environment where there is less parallel processing offered and more demand for CPU time.
2) What is the size of each row - i.e. how many columns and what data is in each column? For example, a row of 100 numbers will copy faster than a row of 100 long text columns.
I have to wonder why your are copying rows? Duplicating data is usually not a good idea. There are some use cases for it such as logging changes or for archiving rows.
LinkedIn
Twitter
