>I have other indexes based on my queries, but yes I am worried about page splits while loading. The PK is an integer (identity). Making that the clustered index would eliminate splits on loads, right?

Probably but it would be unlikely that your queries would take advantage of it. In which case, why have it?

The clustered index is your most useful index. You should really try to take advantage of it.

One option to consider is to drop all indexes when you do your load and then recreate them afterwards.

-Mike




>
>Thanks
>
>John
>
>>I suggest that you take a look at the queries that you execute against the table and prioritize them based on frequency and the audience.
>>
>>Are you worried about page/splits while loading?
>>
>>-Mike
>>
>>>I have a single table database. The table holds health claims information (service codes and dates, providers, consumers) from another agency. It is updated once a month by inserting ~100k - 200k new rows (it currently has 10 million). Existing rows are never updated. I am going to move it to another server and am trying to figure out the best clustered index.
>>>
>>>It seems that any column that would be useful for querys would cause massive page splits, since the data is so diverse for any given month. If I used the month/year paid, it would cause no splits since that is the way the data is sent to us. We do query on (between) service begin and end dates. The claims do come in somewhat in order by those dates.
>>>What does anyone suggest?
>>>
>>>TIA
>>>
>>>John