Microsoft® SQL Server™ version 7.0 does not support custom sort orders. Defining a self-consistent set of sorting rules for a character set can be difficult, and the introduction of Unicode data, with its exponentially larger set of characters, makes this even more challenging.
It is recommended that you use Unicode data types when possible in SQL Server 7.0. In addition to offering an expanded range of characters of Unicode data, a Unicode collation may be compatible with your SQL Server 6.x custom sort order. Unicode collations are equivalent to Microsoft Windows NT® collations for the same locale, and these vary from the SQL Server 6.x sort orders in some instances. Also, the case, accent, width, and kana-sensitivity can be changed for each collation.
If you use a custom sort order with your SQL Server 6.x installation, evaluate what benefit you get from the custom sort order that you do not get from a SQL Server 7.0 sort order and Unicode collation. You may be able to switch without any complications. If the custom sort order presents important benefits, you may be able to compensate programmatically in your applications.
The SQL Server Upgrade Wizard can usually upgrade a custom sort order to the nearest equivalent. However, upgrading from a custom sort order can cause problems if either of these two conditions are true:
If two strings compare as not equal in the custom sort order but as equal in the SQL Server 7.0 sort order, you may violate a uniqueness constraint. If the two strings exist in a SQL Server 6.x database in a unique index, the upgrade will fail.
The sort order may affect how well data is presented, but most applications do not depend on a particular sort order in order to function correctly. An application that fails without the custom sort order in place must be changed to programmatically account for the new sort order.
| Character Set | Unicode Collation |
| Sort Order |