Comments on: VARCHAR (MAX), NVARCHAR (MAX) and NTEXT Data Types

By: anil

anil — Sun, 10 Mar 2013 10:43:19 +0000

Thanks,

By: Pierre

Pierre — Tue, 26 Feb 2013 13:49:13 +0000

sometimes I create nvarchar(max) columns but I index on a calculated column of the first N positions: fits many query needs while matching the size of a string in the application – if you say nvarchar(200), you can be sure you will need 205 one of these days etc

By: Sean

Sean — Thu, 24 Jan 2013 16:16:41 +0000

Perfect article!!! Thank you!

By: Sean

Sean — Thu, 24 Jan 2013 16:16:17 +0000

Very intuitive article, thank you!

By: Sunny

Sunny — Mon, 21 Jan 2013 19:59:35 +0000

Nice article.

By: Andre Ranieri

Andre Ranieri — Mon, 21 Jan 2013 14:24:38 +0000

Jeff:

Good point. Michelle Ufford at GoDaddy wrote a very good blog on “right sizing” and the impact it had on data performance on tables containing billions of rows.

http://inside.godaddy.com/scaling-database-data-types/

By: Aquila

Aquila — Mon, 21 Jan 2013 12:44:16 +0000

Great article, nice and easy to understand for a any novice!!!

By: Mahesh Mohandas

Mahesh Mohandas — Mon, 21 Jan 2013 11:52:20 +0000

Nice one !!!

By: Jeff Moden

Jeff Moden — Mon, 21 Jan 2013 11:30:47 +0000

I have to tell you that if I ever caught a developer making a Student_Name column using any form of blob, I’d have to chase them until their heels smoked.

Perhaps a more appropriate article would be on the extreme advantages of “right sizing”. Just having a BLOB in a table prevents ONLINE indexing of the clustered index in the Enterprise Edition even if the column IS NOT one of the key columns.

By: Philip

Philip — Mon, 21 Jan 2013 05:29:26 +0000

But what about fragmentation? Text is a consistent size (a pointer to another location), and thus fragmentation is lower than when using Varchar, assuming you will never want to index the Text. Is this correct?