I would definitely build clustered indexes on 1GB+ heap tables and would seriously consider applying compression to those indexes.

The SQL Guy @ blogspot[/url]

@SeanPearceSQL

About Me[/url]