Sergiy - Tuesday, March 6, 2018 5:39 PM
That's why it actually does matter. If you have a 100% Fill Factor on a table (clustered index) or non-clustered index that suffers (like Joe Torre said above) out of order inserts or expansive updates to variable width columns and you intend to make the mistake of defragging the index based only on the percent of fragmentation, you'll enter the wonderful world of blocking due to massive page splits right after you rebuild the index. It also has the nasty effect of bloating your log file.
Heh... of course, if you do like the folks did before me where I work and make denormailized tables with 147 columns with a row width of more than 4 KBytes, then, yeah... there's nothing Fill Factor is going do to save you.
Agreed on the 0 Fill Factor. I look at it and 100 as a marker for a probable mistake.
--Jeff Moden
Change is inevitable... Change for the better is not.