GilaMonster (8/26/2016)

---

Jeff Moden (8/26/2016)

---

Grant Fritchey (6/9/2015)

---

There are places where partitioning can, and does, improve performance.

Apologies for the year late response and I hate to be contrary but everywhere I've seen someone make that claim by example, I've seen others beat it with a correctly indexed monolithic table and correctly written code for the problem at hand. I've also beaten a couple of examples myself and it's actually not that difficult.

I've seen it as well (my howlers article on Simple Talk showed it), but in my experience it hasn't been significant enough an improvement that I'm willing to implement something as complex as partitioning to get.

Grant Fritchey (8/26/2016)

---

GilaMonster (8/26/2016)

---

Jeff Moden (8/26/2016)

---

Grant Fritchey (6/9/2015)

---

There are places where partitioning can, and does, improve performance.

Apologies for the year late response and I hate to be contrary but everywhere I've seen someone make that claim by example, I've seen others beat it with a correctly indexed monolithic table and correctly written code for the problem at hand. I've also beaten a couple of examples myself and it's actually not that difficult.

I've seen it as well (my howlers article on Simple Talk showed it), but in my experience it hasn't been significant enough an improvement that I'm willing to implement something as complex as partitioning to get.

Oh believe me, I think the cases where it can work are pretty darned rare, but they do exist. Always and forever (until the technology changes anyway), partitioning is about data management. I say that over and over again. I just know that, in some cases, it does improve performance. I suppose even saying that is opening the door a crack which will cause an entire herd of horses to stampede through it, because everyone always believes that they, and they alone, are the single unique exception to the rule, any rule, all rules. However, it's still true. It's just exceedingly rare.