db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rick Hillegas <Richard.Hille...@Sun.COM>
Subject Re: Collation feature discussion
Date Fri, 30 Mar 2007 20:07:48 GMT
Thanks, Mamta, I think this is a good solution. No doubt there will be 
lots of fun verifying that the current schema changes correctly so that 
the metadata queries run--and in explaining this behavior in the user 
guides! Thanks again for volunteering to build this very useful feature.

Regards,
-Rick

Mamta Satoor wrote:
> Rick, you and Dan have the same concern about the dynamic nature of 
> current schema's character set association with string literal. The 
> main reason behind my thinking was that I didn't want the metadata 
> queries to break in the absence of COLLATE clause. But I had a quick 
> off the list talk with Dan about metadata queries and user queries and 
> he had an excellent solution to the problem. He recommended using the 
> CAST function which will be able to provide the correct character set 
> association in metadata queries and user queries going against system 
> tables.
>  
> So, the string literal will always have the USER character set 
> associated with them (and not current schema's character set).
>  
> The metadata queries which do system character column comparison 
> against string literal can be rewritten as syscharcol = 
> CAST(string_literal as varchar(128)). Since CAST result character 
> string type will pick up the current schema's character set, the 
> CAST's character set will become UCS_BASIC. This way, the metadata 
> queries will work fine because we have the same character set and 
> collation associated on each side of the comparison.
>  
> The user quries accessing system character columns can also use this 
> CAST to their benefit. eg CAST(syscharcol as varchar(128)) = 
> user_char_col. Both the sides of the comparison will have USER 
> character set associated with them.
>  
> So, the only change I am suggesting from my mail on March 29th is that 
> string literal will always have the USER character set associated with 
> them.
>  
> In addition, the answer Dan's question below is yes.
> "Does 3) cover JDBC parameters as well, (ie. ?) where the type of the 
> parameter is a character type?"
>  
> If this covers everything with no further issues, then I will do one 
> final write up on collation derivation and collation type and put it 
> on the wiki page 
> http://wiki.apache.org/db-derby/BuiltInLanguageBasedOrderingDERBY-1478 
>  
> Thanks to everyone for putting in energy to help me arrive on this 
> final writeup step of collation derivation.
> Mamta
>  
> On 3/30/07, *Rick Hillegas* <Richard.Hillegas@sun.com 
> <mailto:Richard.Hillegas@sun.com>> wrote:
>
>     Mamta Satoor wrote:
>     > lots of thoughtful analysis...
>     >
>     >
>     > The character set specification for string literals is not as well
>     > defined as for <column definitions> but my proposal here will work
>     > within SQL spec boundaries. SQL spec Section 5.3<literal>,
>     Syntax Rule
>     > 14b says that if the character set is not specified for character
>     > string literal, then character string literal's character set
>     will be
>     > the character set of the SQL-client module. Derby does not implement
>     > SQL-client module, but definition of SQL-client module in
>     Section 13.1
>     > says that SQL-client module definition has mandatory <module name
>     > clause> which is defined in Section 13.2 <module name clause>. The
>     > Syntax Rule 4 in this section says that if a character set is not
>     > specified for the SQL-client module, then it's character set is
>     > implementation-defined. I think we can use this
>     implementation-defined
>     > character set for a SQL-client module to our advantage. We can
>     define
>     > Derby's implementation-defined character set for SQL-client
>     module as
>     > current schema's character set and hence the current schema's
>     > character set will become string literal's character set. Hence, if
>     > the string literal reference is made while the current schema is
>     user
>     > schema, then the character set associated with string literal
>     will be
>     > USER and if the string literal reference is made while the current
>     > schema is system schema, then the character set associated with the
>     > string literal will be SQL_IDENTIFER. Going this path will also make
>     > sure that our metadata queries don't break.
>     >
>     I'm a bit uneasy about this interpretation. It seems to me that the
>     string literal has a default character set and collation. Those are
>     determined at the database level. They can be overridden at the
>     session
>     level and the client-module level. I think your interpretation of
>     13.2/SR4 is a bit adventurous. To me, 13.2/SR4 says that there is a
>     single, implementation-defined implicit character set for a
>     client-module and not a number of different implicit character sets
>     which shift based on your current schema.
>
>     Regards,
>     -Rick
>
>


Mime
View raw message