Should I use != or <> for not equal in T-SQL?

840
votes

I have seen SQL that uses both != and <> for not equal. What is the preferred syntax and why?

I like !=, because <> reminds me of Visual Basic.

14
sqlsql-servertsql
See also: stackoverflow.com/questions/7884/… – Dinah
Portability of code. If your requirements are easily met by ANSI SQL, then its better to use it. You can use the same code in all DB's. Eg. An SQL book author who wants to illustrate basic SQL using sample code. – Steam
I'd like to add an example where having only ANSI SQL code can be a problem - Standard SQL supports the options NULLS FIRST and NULLS LAST to control how NULLs sort, but T-SQL doesn’t support this option. – Steam
There no need for reopening. The marked question is a duplicate, just extended by yet one more option, NOT (A = B). – TLama
Because it reminds you of Visual Basic is kind of a bad reason. This question can certainly do without that opinion in it. A better reason would be like the reason in one of the answers where SQL is being stored in XML. Not sure why one would want to store SQL in XML, but it's a better reason nonetheless. – Michael Z.

14 Answers

571
votes

Technically they function the same if you’re using SQL Server AKA T-SQL. If you're using it in stored procedures there is no performance reason to use one over the other. It then comes down to personal preference. I prefer to use <> as it is ANSI compliant.

You can find links to the various ANSI standards at...

http://en.wikipedia.org/wiki/SQL

750
votes

Most databases support != (popular programming languages) and <> (ANSI).

Databases that support both != and <>:

Databases that support the ANSI standard operator, exclusively:

  • IBM DB2 UDB 9.5: <>
  • Microsoft Access 2010: <>
115
votes

'<>' is from the SQL-92 standard and '!=' is a proprietary T-SQL operator. It's available in other databases as well, but since it isn't standard you have to take it on a case-by-case basis.

In most cases, you'll know what database you're connecting to so this isn't really an issue. At worst you might have to do a search and replace in your SQL.

44
votes

The ANSI SQL Standard defines <> as the "not equal to" operator,

http://www.contrib.andrew.cmu.edu/~shadow/sql/sql1992.txt (5.2 <token> and <separator>)

There is no != operator according to the ANSI/SQL 92 standard.

22
votes

It seems that Microsoft themselves prefer <> to != as evidenced in their table constraints. I personally prefer using != because I clearly read that as "not equal", but if you enter [field1 != field2] and save it as a constrait, the next time you query it, it will show up as [field1 <> field2]. This says to me that the correct way to do it is <>.

15
votes

!=, despite being non-ANSI, is more in the true spirit of SQL as a readable language. It screams not equal. <> says it's to me (less than, greater than) which is just weird. I know the intention is that it's either less than or greater than hence not equal, but that's a really complicated way of saying something really simple.

I've just had to take some long SQL queries and place them lovingly into an XML file for a whole bunch of stupid reasons I won't go into.

Suffice to say XML is not down with <> at all and I had to change them to != and check myself before I riggedy wrecked myself.

12
votes

You can use whichever you like in T-SQL. The documentation says they both function the same way. I prefer !=, because it reads "not equal" to my (C/C++/C# based) mind, but database gurus seem to prefer <>.

10
votes

I understand that the C syntax != is in SQL Server due to its Unix heritage (back in the Sybase SQL Server days, pre Microsoft SQL Server 6.5).

7
votes

One alternative would be to use the NULLIF operator other than <> or != which returns NULL if the two arguments are equal NULLIF in Microsoft Docs. So I believe WHERE clause can be modified for <> and != as follows:

NULLIF(arg1, arg2) IS NOT NULL

As I found that, using <> and != doesn't work for date in some cases. Hence using the above expression does the needful.

2
votes

I preferred using != instead of <> because sometimes I use the <s></s> syntax to write SQL commands. Using != is more handy to avoid syntax errors in this case.

-6
votes

They are both accepted in T-SQL. However, it seems that using <> works a lot faster than !=. I just ran a complex query that was using !=, and it took about 16 seconds on average to run. I changed those to <> and the query now takes about 4 seconds on average to run. That's a huge improvement!

-14
votes

Although they function the same way, != means exactly "not equal to", while <> means greater than and less than the value stored.

Consider >= or <=, and this will make sense when factoring in your indexes to queries... <> will run faster in some cases (with the right index), but in some other cases (index free) they will run just the same.

This also depends on how your databases system reads the values != and <>. The database provider may just shortcut it and make them function the same, so there isn't any benefit either way.PostgreSQL and SQL Server do not shortcut this; it is read as it appears above.