Hi,
I'm an Oracle 9i/10g DBA with quite a few years experience, but I'm new to XML and dealing with it in database terms. I've been given a project that entails pulling XML values out of a file (or 100's of them) and storing them in the database so that they are searchable by end-users. The project is classified as secret so I'm unable to upload the specific XML or any info relating to the structire of the XML or the table I will use to insert the values into - sorry!! So, I've created an XML file with a similar structure to help people understand my predicament.
The end-users only need to search on a subset of the total amount of columns from the table I'll insert data into, although the XML file has a lot more, so I dont need to store the other values - but I will need to store the name of the XML file (or a pointer to it so I know what XML file a particular set of values belong to) in another column of the table along with its associated values.
I've been using the XMLTABLE function with some degree of success, although I had better succes using the XMLSEQUENCE function. However, I found out this is deprecated in 10g and replaced with XMLTABLE, so I guess it's better if I use this in case we ever need to upgrade to 11g.
The main problem I've been having is that some elements in the XML files have multiple values for the one record when all the other records are the same. In terms of storing this in the database, I guess it would mean inserting multiple rows in the table for each element where the value differs. Here is a dumbed down XML file similar to what I've got along with the other SQL I've used:
+<?xml version="1.0" encoding="UTF-8"?>+
+<House>+
+<Warehouse>+
+<WarehouseId>1</WarehouseId>+
+<WarehouseName>+
+<Town>Southlake</Town>+
+<State>Texas</State>+
+</WarehouseName>+
+<Building>Owned</Building>+
+<Area>25000</Area>+
+<Docks>2</Docks>+
+<DockType>Rear load</DockType>+
+<WaterAccess>true</WaterAccess>+
+<RailAccess>N</RailAccess>+
+<Parking>Street</Parking>+
+<VClearance>10</VClearance>+
+</Warehouse>+
+<Warehouse>+
+<WarehouseId>2</WarehouseId>+
+<WarehouseName>+
+<Town>Poole</Town>+
+<State>Dorset</State>+
+</WarehouseName>+
+<WarehouseName>+
+<Town>Solihull</Town>+
+<County>West Midlands</State>+
+</WarehouseName>+
+<Building>Owned</Building>+
+<Area>40000</Area>+
+<Docks>5</Docks>+
+<DockType>Rear load</DockType>+
+<WaterAccess>true</WaterAccess>+
+<RailAccess>N</RailAccess>+
+<Parking>Bay</Parking>+
+<VClearance>10</VClearance>+
+</Warehouse>+
+<Warehouse>+
+<WarehouseId>3</WarehouseId>+
+<WarehouseName>+
+<Town>Fleet</Town>+
+<County>Hampshire</County>+
+</WarehouseName>+
+<Building>Owned</Building>+
+<Area>10000</Area>+
+<Docks>1</Docks>+
+<DockType>Side load</DockType>+
+<WaterAccess>false</WaterAccess>+
+<RailAccess>N</RailAccess>+
+<Parking>Bay</Parking>+
+<VClearance>20</VClearance>+
+</Warehouse>+
+</House>+
CREATE TABLE xmltest OF XMLTYPE;
INSERT INTO xmltest
VALUES(xmltype(bfilename('XML_DIR', 'test.xml'), nls_charset_id('AL32UTF8')));
Consequently, I need to...
1) Retrieve the results from the XML file for all 3 warehouses where multiple values for the same sub-element are shown as 2 rowsthe result set. (I am guessing there will be 4 rows returned as warehouse sub-2 has 2 different elements for <WarehouseName>.
2) Build a case statement into the query so that regardless of the sub-element name (i.e State or County), it is returned into the 1 column, for instance County.
So, if I run a query similar to the following...
select y.WarehouseId, y.Town, y.County, y.Area
from xmltest x, xmltable('/House/Warehouse' .......
I would like to get results back like this...
ID Town County Area
1 Southlake Texas 25000
2 Poole Dorset 40000
2 Solihull West Midlands 40000
3 Fleet hampshire 10000
Sorry for the non-formatting but I hope this all makessense to someone out there with what I'm trying to do.
I appreciate any help whatsoever because, as i said before, I'm totally new to XML and trying to read the vast amount of information there is out there on XML is all a bit daunting.
Many thanks in advance,
Shaun.