I'm bringing in a single line of machine generated metadata and then using calculations to parse the pieces of data i need out into separate fields. But as this is time based data I have hundreds of thousands of records.
What are the best settings to keep this moving as fast as possible? Indexing the calculations or not?
I can work with it slow but would like to make it as speedy as possible.
Indexed, stored calculations will evaluate more quickly, (and you can search and sort on them much faster), but produce much slower import speeds if this data is imported from another source--so it's a trade off here depending on the design and needs of your database.