distinct count with multiple partitions?
we analyze of our data using distinct measures. of underlying sources - i.e. daily page views heavy , have millions of rows day. because of i've created daily partitions process incremental data that's arrived. however, i'm curious - how distinct count perform when needs rollup on multiple partitions?
ie it's dec 25th, , have 25 unique partitions each day of december far. internally, how distinct measures correctly accumulate unique instances of measure? there significant performance concerns aware of?
ie it's dec 25th, , have 25 unique partitions each day of december far. internally, how distinct measures correctly accumulate unique instances of measure? there significant performance concerns aware of?
distinct count measure work fine multiple partitions. keeps distinct count values inside partitions, , therefore can correctly aggregate across them. partition per day should ok too.
SQL Server > SQL Server Analysis Services
Comments
Post a Comment