In SQL Server, your ranges also need to ACCENT-uate the positives!

You might have started to sense a pattern developing here with the last few posts in this series:

• In SQL Server, the distance between A and Z is wider than you might think
• In SQL Server, A-Z, A-z, a-Z, and a-z may not mean the same thing
• In SQL Server, the wrong range can make you seem insensitive to one's width! (aka Do my V's look fat?)

That last post pointed out two characters that would be missing if you have a width sensitive collation.

and now this post is going to talk about accent sensitivity and it's impact.

This will add over two dozen characters like the following ones:

• ẑ    U+1e91    LATIN SMALL LETTER Z WITH CIRCUMFLEX
• Ż    U+017b    LATIN CAPITAL LETTER Z WITH DOT ABOVE
• ź    U+017a    LATIN SMALL LETTER Z WITH ACUTE
• ᴢ    U+1d22    LATIN LETTER SMALL CAPITAL Z
• Ȥ    U+0224    LATIN CAPITAL LETTER Z WITH HOOK
• ɀ    U+0240    LATIN SMALL LETTER Z WITH SWASH TAIL

and more! All of which have some sort of "Z-ish" quality about them, none of which will be handled by the ranges defined thus far according to the documentation for SAL Server if your application is run in the context of an accent sensitive collation.

Oops.

You get the point - this odd combination of actual linguistic characters used in language, phonetic symbols, and circled/parenthesize letters have a fundamental identity that users would expect to sort somewhere not far from the Z by default.

SQL Server does that well -- how about your application with your LIKE range? :-)

And just in case you thought I was done, it gets worse tomorrow....

This post brought to you by ⒵ (U+24b5, aka PARENTHESIZED LATIN SMALL LETTER Z, the last Z-like thing in the Latin script....)