You can read the current date and time in SQL Server in eight ways, and they differ in precision and time zone. This post runs them side by side, shows that each one is evaluated once per statement, and picks the right one for defaults, audit columns and reports.
SQL SERVER – Find Length of Text Field
LEN works for VARCHAR but not for TEXT. To find the length of text field data, use DATALENGTH(textfield). A one line example shows how.
SQLAuthority.com News – Journey to SQL Authority Milestone of SQL Server
My 200th post, a thank you to my readers and a look at the Journey to SQL Authority milestones: interview questions, coding standards, jobs, humor and more.
SQL SERVER – Delay Function – WAITFOR clause – Delay Execution of Commands
The WAITFOR clause delays a batch, stored procedure or transaction until a set time or for an interval. I use it daily after restores. DELAY and TIME examples.
SQL SERVER – De-fragmentation of Database at Operating System to Improve Performance
Why does de-fragmentation of database files at the operating system level skip MDF files? SQL Server has them open, so stop the services first.
SQL SERVER – 2005 – UDF – User Defined Function to Strip HTML – Parse HTML – No Regular Expression
A simple SQL Server function to parse HTML and keep only the text, with no regular expressions needed. Copy the UDF and try it on your own data.
SQL SERVER – sp_HelpText for sp_HelpText – Puzzle
What happens when you run sp_HelpText on itself? It prints its own code. Plus tips for long lines, encrypted objects and searching object code.
SUM Integer Values: Cast Before the Aggregate
SUM integer values can exceed int even when each row fits. Cast each input to bigint before aggregation and keep NULL and empty-set behavior visible.
SQL SERVER – Easy Sequence of SELECT FROM JOIN WHERE GROUP BY HAVING ORDER BY
The sequence of SELECT, FROM, JOIN, WHERE, GROUP BY, HAVING and ORDER BY that I gave my new hires. Since then they ask me far fewer questions.
SQL SERVER – Explanation SQL SERVER Hash Join
How a Hash Join works in two phases, build and probe, why it suits large data sets without indexes, and why it needs an equijoin predicate.