db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Army <qoz...@gmail.com>
Subject Re: XML type description - possible changes?
Date Tue, 29 Aug 2006 17:18:38 GMT
Daniel John Debrunner wrote:

<snip old XML type text>
> 
> That seems wrong to me, the use of "store Unicode character-based", any
> implementation details are not important here. Also the use of "large"
> is not relevant, what about small XML documents?

I have to admit, I'm guilty of copy-paste here.  The current documentation for 
the CLOB type has the following sentence, which is the root of the questions 
raised in this email:

<begin quote>

A CLOB (character large object) value can be up to 2,147,483,647 characters 
long. A CLOB is used to store unicode character-based data, such as large 
documents in any character set.

<end quote>

I copied that as my starting point and failed to clean it up...sorry.

> Looking at the other type definitions it seems something like this would
> match, though it seems unwieldy.
> 
> "XML provides for storage of Extensible Markup Language (XML) documents,
> that conform to the SQL/XML definition of a well-formed
> XML(DOCUMENT(ANY)) value."

I agree, I think this is a better wording.  My one reservation is that the XML 
type can also be used transiently for XML values that are not well-formed 
documents.  In particular, the XMLQUERY operator returns a value of type XML 
that is not guaranteed to be XML(DOCUMENT(ANY)).

So while I agree with the new text proposal, I think it'd be good (or at least, 
more accurate) to mention that XML is not restricted to "storage of ... 
well-formed XML(DOCUMENT(ANY)) values"; it also provides for transient use of 
XML(SEQUENCE) values, which may or may not be well-formed XML(DOCUMENT(ANY)) values.

> Also I'm not sure about the character limitation on the XML value, is
> that the limit on the input string to the document, or its final
> storage?

My impression is that this is the limit on the final storage--and that limit is, 
so far as I know, a Derby limitation, not a SQL/XML one.  At least, that's what 
I assumed based on the fact that CLOBs in Derby have the same hard limit.  I 
admit I could be wrong, here...please correct me if needed.

> Not sure describing the final stored limit in terms of
> characters makes sense here.

Agreed.  Especially since it's possible that the way in which Derby stores XML 
could change in the future for better performance and/or extended functionality 
(XML indexing comes to mind).  So no, we probably should not be describing the 
stored limit in terms of characters.  I think that line could be removed entirely.

Should I add this info to the Reference Guide documentation wiki?

http://wiki.apache.org/db-derby/ReferenceManualTenTwo#sqlreference

Or would you like to? :)

Thanks for taking the time to buddy test the XML doc and features; I appreciate 
the feedback...

Army


Mime
View raw message