Lists Home |
Date Index |
At 03:17 PM 3/27/2003 +0000, Neile Bermudes wrote:
>What I would like to do: first, export all the information in Quark into a
>SQL database, having exactly the same information in both. Then, as
>details change, just update the database. Once a year, when we print out
>the Staff Directory booklet, I would like to then like to get the updated
>information into Quark so it can be printed.
First, if you ping me in about three or four weeks, I may be able to give
you an early draft of a paper that compares doing this by hand using the
DOM and SQL queries, proprietary vendor extensions from major database
vendors, using SQL/XML to generate XML from relational data, and XQuery on
Does your solution need to work for multiple database vendors, or is there
one particular brand of database that you care about? This makes a big
difference. How do you feel about tedious programming that may not run all
that efficiently - is this a process that just has to run once a year,
where the data is not all that complex? Doing it by hand may not be too bad.
If your data isn't too complex, you can do this by hand. This does get
Every major database vendor has their own proprietary approaches that solve
this problem. Some of these solutions are better than others, most are a
bit ad hoc.
If you want a solution that works for more than one database vendor, there
are two standards to consider: XQuery and SQL/XML.
SQL/XML is a set of proposed extensions to the SQL standard that has been
implemented by Oracle and IBM, and is expected to be adopted in 2003. My
company has a cross-database implementation based on an early version of
SQL/XML, and will soon have a new release that conforms to the latest
drafts. SQL/XML allows you to write SQL queries that return XML as their
results. Suppose you have a table like this:
EmpId LastName FirstName
----- ---------- -------------------------
1 Marshall Marc
2 Ayers Brian
3 Simpson Joanna
4 O'Donnel Gavin
You can create XML from this data using a query like this:
xmlelement ( name "employee",
xmlattributes (e.EmpId as "id"),
xmlelement ( name "names",
xmlelement ( name "first", e.FirstName),
xmlelement ( name "last", e.LastName)
from Employees e
The result of the above query would be:
This is pretty easy for SQL programmers to learn. It may not be obvious
from this simple query, but when queries more complex, involving grouping,
multiple joins, and mapping the tabular representation of SQL into various
hierarchies, SQL/XML can be a significant timesaver - and more efficient at
run-time than programming it yourself.
My company, DataDirect Technologies, has extended SQL/XML with updates that
use XPath to identify the data to be used for updating a relational
database. You could use this to insert the data from your XML files with a
simple Update statement. Our current syntax is not quite conformant with
the standard, but in the second quarter of this year, we will be releasing
a version that is.
The other standard you should care about is XQuery, which takes XML as
input and output. Many vendors are supplying an XML view of a relational
database. I think this will be a really good way to go fairly soon, as
XQuery implementations on relational databases become more mature. Most
relational database vendors will be offering this, as well as third
parties, including my own company.
I'm out of time for this right now, but here are some resources you may be
Ron Bourret's XML and Databases page:
SQL/XML: Two readable articles:
SQL/XML: The current drafts of the proposed standard:
Hope that helps!