Question

I’m trying to store 46,000 objects in Sterling and it’s taking 3 minutes.

Yes I know it’s a lot but this is data provided by the customer and could end up being a lot more.

I’m guessing each time I save a new object it is looking up the key to see if the object has already been stored.

Is there any way to bypass this and tell sterling to just insert?

Any other ideas?

Answer 1

Without knowing anything about your data structure it’s a bit difficult to recommend ways in which you could improve performance, however:

• The fewer indexes you create for your data tables the fewer indexes there are to create when your data is persisted. You should look carefully at which indexes you need for your data read scenarios.
• The more data relationships there are, the more metadata there is to create at write time. You may be able to simplify the data structures and combine classes.
• Sheer volume of data sounds like your biggest problem. I’ve experienced similar problems before with trying to persist large volumes of GPS data. The problem there is that I was trying to write a lot of relatively small amounts of data related to a single piece of data in another table. I managed to resolve this by consolidating the GPS data into a single string and persisting it as a field with the main record. This offloaded a lot of the read/write time into a significantly smaller amount of time for rehydrating the data when it was actually needed.

I woudl definitely recommend reachign out to Jeremy and the Sterling team via the CodePlex site if none of the above help.