Unicode conversion, database woes (Delphi 2007 to XE2) - sql-server

Currently, I am in the process of updating all of our Delphi 2007 code base to Delphi XE2. The biggest consideration is the ANSI to Unicode conversion, which we've dealt with by re-defining all base types (char/string) to ANSI types (ansichar/ansistring). This has worked in many of our programs, until I started working with the database.
The problem started when I converted a program that stores information read from a file into an SQL Server 2008 database. Suddenly simple queries that used a string to locate data would fail, such as:
SELECT id FROM table WHERE name = 'something'
The name field is a varchar. I found that I was able to complete the query successfully by prefixing the string name with an N. I was under the impression that varchar could only store ANSI characters, but it appears to be storing Unicode?
Some more information: the name field in Delphi is string[13], but I've tried dropping the [13]. The database collation is SQL_Latin1_General_CP1_CI_AS. We use ADO to interface with the database. The connection information is stored in the ODBC Administrator.
NOTE: I've solved my actual problem thanks to a bit of direction from Panagiotis. The name we read from our map file is an array[1..24] of AnsiChar. This value was being implicitly converted to string[13], which was including null characters. So a name with 5 characters was really being stored as the 5 characters + 8 null characters in the database.

varchar fields do NOT store Unicode characters. They store ASCII values in the codepage specified by the field's collation. SQL Server will try to convert characters to the correct codepage when you try to store Unicode or data from a different codepage. You can disable this feature but the best option is to avoid the whole mess by using nvarchar fields and UnicodeString in your application.
You mention that you changes all character types to ANSI, not UNICODE types in your application. If you want to use UNICODE you should be using a UNICODE type like UnicodeString. Otherwise your values will be converted to ANSI when they are sent to your server. This conversion is done by your code when you create the AnsiString that is sent to the server.
BTW, your select statement stores an ASCII value in the field. You have to prepend the value with N if you want to store it as a unicode value, eg.g
SELECT id FROM table WHERE name = N'something'
Even this will not guarantee that your data will reach the server in a Unicode form. If you store the statement in an AnsiString the entire statement is converted to ANSI before it is sent to the server. If your app makes a wrong conversion, you will end up with mangled data on the server.
The solution is very simple, just use parameterized statements to pass unicode values as unicode parameters and store them in NVarchar fields. It is much faster, avoids all conversion errors and prevents SQL injection attacks.

Related

Microsoft SQL Server database character conversion

I have a database imported that contains Arabic characters but are displayed as question marks. Through some searching I found that the column data types should be nvarchar to support unicode (they were varchar). I changed one of the columns that contains those characters to nvarchar but the data is still displayed as "??". How can I change the existing values to become unicode and display correctly?
You cannot just simply change the datatype to nvarchar - that won't bring back the data since it's already been "destroyed" by having been converted to a non-Unicode format.
You need to use nvarchar and then you need to insert (or update) the data in such a way that doesn't convert it back to ANSI codes.
If you use T-SQL to insert that Unicode code, make sure to use the N'...' prefix:
INSERT INTO dbo.YourTable(NvarcharCol)
VALUES (N'nvarchar-value')
From a front-end language like C# or PHP or Ruby, make sure to use Unicode strings - .NET (C# and VB.NET) does that automatically. When using queries with parameters, make sure to specify Unicode string types for those relevant parameters.
You need different collation.
Read more here: https://msdn.microsoft.com/en-us/library/ms143508.aspx

Difference between NVARCHAR in Oracle and SQL Server?

We are migrating some data from sql server to oracle. For columns defined as NVARCHAR in SQL server we started creating NVARCHAR columns in Oracle thinking them to be similar..But it looks like they are not.
I have read couple of posts on stackoverflow and want to confirm my findings.
Oracle VARCHAR2 already supports unicode if the database character set is say AL32UTF8 (which is true for our case).
SQLServer VARCHAR does not support unicode. SQLServer explicitly requires columns to be in NCHAR/NVARCHAR type to store data in unicode (specifically in the 2 byte UCS-2 format)..
Hence would it be correct to say that SQL Server NVARCHAR columns can/should be migrated as Oracle VARCHAR2 columns ?
Yes, if your Oracle database is created using a Unicode character set, an NVARCHAR in SQL Server should be migrated to a VARCHAR2 in Oracle. In Oracle, the NVARCHAR data type exists to allow applications to store data using a Unicode character set when the database character set does not support Unicode.
One thing to be aware of in migrating, however, is character length semantics. In SQL Server, a NVARCHAR(20) allocates space for 20 characters which requires up to 40 bytes in UCS-2. In Oracle, by default, a VARCHAR2(20) allocates 20 bytes of storage. In the AL32UTF8 character set, that is potentially only enough space for 6 characters though most likely it will handle much more (a single character in AL32UTF8 requires between 1 and 3 bytes. You probably want to declare your Oracle types as VARCHAR2(20 CHAR) which indicates that you want to allocate space for 20 characters regardless of how many bytes that requires. That tends to be much easier to communicate than trying to explain why some 20 character strings are allowed while other 10 character strings are rejected.
You can change the default length semantics at the session level so that any tables you create without specifying any length semantics will use character rather than byte semantics
ALTER SESSION SET nls_length_semantics=CHAR;
That lets you avoid typing CHAR every time you define a new column. It is also possible to set that at a system level but doing so is discouraged by the NLS team-- apparently, not all the scripts Oracle provides have been thoroughly tested against databases where the NLS_LENGTH_SEMANTICS has been changed. And probably very few third-party scripts have been.

Japanese Character in Oracle Database not displayed properly

I recently created an Oracle DB with JA16SJIS Character Set.
And then I try to insert some data include Japanese characters using SQL*Plus running an external SQL file. The file is encoded in Shift-JIS (and I can see Japanese characters properly in the file using notepad++).
Inserting was success but when I select the data (using SQL*Plus), Japanese characters are not displayed properly (like some alphabet characters with some question marks).
Even when I use SQL Developer to view the data, Japanese characters still unreadable.
And I'm using Window 7 Professional SP1, Oracle Database 11g R2, system locale set to Japan as well.
First, you should try to insert some text directly from SQLDeveloper data view. That should work no matter what, so you can use it to check your imports.
Then before you connect with SQL*Plus you must specify what you're going to send by setting or changing the value of environment variable NLS_LANG.
NSL_LANG=ENGLISH_FRANCE.JA16SJIS
The syntax will depend on your OS. The only important part is the last one JA16SJIS which means Shift-Jis as you already know.
You can then connect with SQL*Plus and import your file.
Note that the encoding that you specify must match the encoding of your file but not necessarily the encoding of the base as Oracle will do a conversion if necessary. So you could have your base in UTF8 and it would still work (because UTF8 can hold japanese characters).
In these cases the first thing I do is to have a look at what byte values are stored in the database. You can use the dump function for that.
select dump(<column>) from <table>
If you know what byte values your characters should have you can check if the correct values are in your table.

Unicode Data Type in SQL

I'm new to Microsoft SQL. I'm planning to store text in Microsoft SQL server and there will be special international characters. Is there a "Data Type" specific to Unicode or I'm better encoding my text with a reference to the unicode number (i.e. \u0056)
Use Nvarchar/Nchar (MSDN link). There used to be an Ntext datatype as well, but it's deprecated now in favour of Nvarchar.
The columns take up twice as much space over the non-unicode counterparts (char and varchar).
Then when "manually" inserting into them, use N to indicate it's unicode text:
INSERT INTO MyTable(SomeNvarcharColumn)
VALUES (N'français')
When you say special international characters, what do you mean? If special means they aren't common and just occasional, then the overhead of nvarchar might not make sense in your situation on a table with a very large number of rows or a lot of indexing.
I'm all for using Unicode where appropriate, but understanding when it is appropriate is important.
If you are mixing data with different implied code pages (Japanese and Chinese in same database) or you just want to be forward-looking for internationalization and localization, then you want the column to be Unicode and use nvarchar data type and that's perfectly fine. Unicode is not going to magically solve all sorting problems for you.
If you are know that you will always be storing mainly ASCII but some occasional foreign characters, just store your UTF-8 data or HTML encoded data in varchar. If your data is all in Japanese and code page 932 (or any other single code page), you can still store double-byte characters in varchar, they still take up two bytes. My point is, that when you are already in a DBCS collation, international characters are no longer "special". It's not just the data storage, but any indexes as well as the working set when dealing with such a column in queries and in other dataflows.
And do not make a blanket rule that all character data should be nvarchar - it's a waste for many columns which are codes or identifiers.
Any time you have a column, go through the same questions:
What is the type of data?
What is the range?
Are NULLs allowed?
What is the limit of the size?
Are there any constraints I should apply now to stop bad data getting in from the beginning?
People have had success with using the following code to force Unicode at insert data manipulation.
INSERT INTO <table> (text) values (N'<text here>)
1
Character set features of tables and string inside them are specified for the database and if your database has a Unicode collation, strings inside the tables are unicode. As well for string columns you have to use nvarchar or nchar data types to make them able to store unicode strings. But this feature works if your database has a utf8 or unicode characterset or collation. Read this link for more information. Unicode and SQL Server

How can I recover Unicode data which displays in SQL Server as?

I have a database in SQL Server containing a column which needs to contain Unicode data (it contains user's addresses from all over the world e.g. القاهرة‎ for Cairo)
This column is an nvarchar column with a collation of database default (Latin1_General_CI_AS), but I've noticed data inserted into it via SQL statements containing non English characters and displays as ?????.
The solution seems to be that I wasn't using the n prefix e.g.
INSERT INTO table (address) VALUES ('القاهرة')
Instead of:
INSERT INTO table (address) VALUES (n'القاهرة')
I was under the impression that Unicode would automatically be converted for nvarchar columns and I didn't need this prefix, but this appears to be incorrect.
The problem is I still have some data in this column which appears as ????? in SQL Server Management Studio and I don't know what it is!
Is the data still there but in an incorrect character encoding preventing it from displaying but still salvageable (and if so how can I recover it?), or is it gone for good?
Thanks,
Tom
To find out what SQL Server really stores, use
SELECT CONVERT(VARBINARY(MAX), 'some text')
I just tried this with umlauted characters and Arabic (copied from Wikipedia, I have no idea) both as plain strings and as N'' Unicode strings.
The results are that Arabic non-Unicode strings really end up as question marks (0x3F) in the conversion to VARCHAR.
SSMS sometimes won't display all characters, I just tried what you had and it worked for me, copy and paste it into Word and it might display it corectly
Usually if SSMS can't display it it should be boxes not ?
Try to write a small client that will retrieve these data to a file or web page. Check ALL your code if there are no other inserts or updates that might convertthe data to varchar before storing them in tables.

Resources