Does SQL server expect numbers to be specified with digits from the latin alphabet, e.g.:
0123456789
Is it valid to give SQL Server digits in other alphabets?
Rosetta Stone:
Latin: 01234567890
Arabic: ٠١٢٣٤٥٦٧٨٩
Bengali: ০১২৩৪৫৬৭৮৯
i know that the client (ADO) will convert 8-bit strings to 16-bit unicode strings using the current culture. But the client is also converting numbers to strings using their current culture, e.g.:
SELECT * FROM Inventory
WHERE Quantity > ২৩৪,৭৮
Which throws SQL Server for fits.
i know that the server/database has it's defined code page and locale, but that is for strings.
Will SQL Server interpret numbers using the active (or per-login specified) locale, or must all numeric values be specifid with latin numeral digits?
From what I can tell, T-SQL requires latin digits, and decimal points specified as
.
.Neither ISNUMERIC() nor CAST() can successfully test these digits, so a numeric constant using those characters would not work either.
Allowing a client to pass non-Latin digits sounds dangerously promiscuous (I'm not sure what path your data travels, but there seems to be a potential for SQL injection if user's localized input isn't being tested to be numeric.