Inserting parent /child records from a xml file ...

Question

XML file pasted below:
I have loaded the xml file into a table called xml_demo which has a column xml_data of type xmltype.
The select for parent record is as follows and it works:
INSERT INTO balit_submissions (balitdoc, documentversion, datetime_from,job_id, status,creation_datetime)
(SELECT 'MOL'
,to_number(extract(x.xml_data,'/MolDocument/DocumentVersion/@v'))
,to_date(substr(extract(x.xml_data,'/MolDocument/ValidTimeInterval/@v'),1,16),
'yyyy-mm-dd"T"hh24:mi')
,123456
,'CREATED',
NULL
FROM xml_DEMO x WHERE
existsnode(x.xml_data,'/MolDocument/DocumentIdentification[@v="MOL_20100331_1500_1600"]') = 1)
Having problem creating child records. From this file I expect to create 3 records.
insert into balit_energy_blocks
SO_ID,
DATETIME_FROM,
DIRECTION,
BLOCK_NUMBER,
ENERGY,
LAST_SUBMIT_DATETIME,
PRICE_POUNDS,
PRICE_EUROS,
BALIT_REF,
STATUS,
LAST_EDIT_DATETIME,
MOL_REASON,
ACQUIRING_SO_AREA)
(SELECT 'RTE'
,to_date(substr(extract(x.xml_data,'/MolDocument/ValidTimeInterval/@v'),1,16),
'yyyy-mm-dd"T"hh24:mi')
,DECODE(extract(x.xml_data,'/MolDocument/MolTimeSeries/Direction/@v'),'AO1','Up','Down')
,to_number(substr(extract(x.xml_data,'/MolDocument/MolTimeSeries/ContractIdentification/@v'),19))
,to_number(extract(x.xml_data,'/MolDocument/MolTimeSeries/Period/Interval/EnergyPrice/@v'))
,sysdate
,null -- price pounds
,null -- price euro
,extract(x.xml_data,'/MolDocument/MolTimeSeries/ContractIdentification/@v')
,'SUBMITTED'
,'A96'
,NULL -- acquiring area
,sysdate
FROM xml_DEMO x WHERE
existsnode(x.xml_data,'/MolDocument/DocumentIdentification[@v="MOL_20100331_1500_1600"]') = 1)
For example, there are 3 ContractIdentification tags. Example of 1:
<ContractIdentification v="RTE_20100331_1500_16"/>
I was expecting this select to pluck the last number from this string. In this case 16.
The select was:
to_number(substr(extract(x.xml_data,'/MolDocument/MolTimeSeries/ContractIdentification/@v'),19))
The result I got was:
16RTE_20100331_1500_20NG_20100331_1500_6
All contractident values are concatnated and returns from position 19 onwards.
Can anyone help me to extract the last number from each ContractIdentification tag value and to create the 3 records
Thanks
James Sathiaraj
<?xml version="1.0" encoding="UTF-8"?>
<MolDocument DtdVersion="3" DtdRelease="0">
<DocumentIdentification v="MOL_20100331_1500_1600"/>
<DocumentVersion v="1"/>
<DocumentType v="A43"/>
<SenderIdentification codingScheme="A01" v="17X100Z100Z0001H"/>
<SenderRole v="A35"/>
<ReceiverIdentification codingScheme="A01" v="10XFR-RTE------Q"/>
<ReceiverRole v="A04"/>
<CreationDateTime v="2010-03-31T14:10:00Z"/>
<ValidTimeInterval v="2010-03-31T15:00Z/2010-03-31T16:00Z"/>
<Domain codingScheme="A01" v="10YDOM-1001A001A"/>
<MolTimeSeries>
<ContractIdentification v="RTE_20100331_1500_16"/>
<ResourceProvider codingScheme="A01" v="10XFR-RTE------Q"/>
<AcquiringArea codingScheme="A01" v="17Y100Z100Z00013"/>
<ConnectingArea codingScheme="A01" v="10YFR-RTE------C"/>
<AuctionIdentification v="AUCTION_20100331_1500_1600"/>
<BusinessType v="A10"/>
<BidTimeInterval v="2010-03-31T15:00Z/2010-03-31T16:00Z"/>
<MeasureUnitQuantity v="MAW"/>
<Currency v="EUR"/>
<MeasureUnitPrice v="MWH"/>
<Direction v="A02"/>
<MinimumActivationQuantity v="50"/>
<Status v="A06"/>
<Period>
<TimeInterval v="2010-03-31T15:00Z/2010-03-31T16:00Z"/>
<Resolution v="PT60M"/>
<Interval>
<Pos v="1"/>
<Qty v="50"/>
<EnergyPrice v="50.45"/>
</Interval>
</Period>
</MolTimeSeries>
<MolTimeSeries>
<ContractIdentification v="RTE_20100331_1500_20"/>
<ResourceProvider codingScheme="A01" v="10XFR-RTE------Q"/>
<AcquiringArea codingScheme="A01" v="17Y100Z100Z00013"/>
<ConnectingArea codingScheme="A01" v="10YFR-RTE------C"/>
<AuctionIdentification v="AUCTION_20100331_1500_1600"/>
<BusinessType v="A10"/>
<BidTimeInterval v="2010-03-31T15:00Z/2010-03-31T16:00Z"/>
<MeasureUnitQuantity v="MAW"/>
<Currency v="EUR"/>
<MeasureUnitPrice v="MWH"/>
<Direction v="A02"/>
<MinimumActivationQuantity v="50"/>
<Status v="A06"/>
<Period>
<TimeInterval v="2010-03-31T15:00Z/2010-03-31T16:00Z"/>
<Resolution v="PT60M"/>
<Interval>
<Pos v="1"/>
<Qty v="50"/>
<EnergyPrice v="50.48"/>
</Interval>
</Period>
</MolTimeSeries>
<MolTimeSeries>
<ContractIdentification v="NG_20100331_1500_6"/>
<ResourceProvider codingScheme="A01" v="10X1001A1001A515"/>
<AcquiringArea codingScheme="A01" v="17Y100Z100Z00013"/>
<ConnectingArea codingScheme="A01" v="10YGB----------A"/>
<AuctionIdentification v="AUCTION_20100331_1500_1600"/>
<BusinessType v="A10"/>
<BidTimeInterval v="2010-03-31T15:00Z/2010-03-31T16:00Z"/>
<MeasureUnitQuantity v="MAW"/>
<Currency v="EUR"/>
<MeasureUnitPrice v="MWH"/>
<Direction v="A01"/>
<MinimumActivationQuantity v="50"/>
<Status v="A06"/>
<Period>
<TimeInterval v="2010-03-31T15:00Z/2010-03-31T16:00Z"/>
<Resolution v="PT60M"/>
<Interval>
<Pos v="1"/>
<Qty v="50"/>
<EnergyPrice v="17.0"/>
</Interval>
</Period>
</MolTimeSeries>
</MolDocument>

Answer 1

Hi,
The result I got was:
16RTE_20100331_1500_20NG_20100331_1500_6In the query you tried, you access a single record so you can't expect to get three rows "magically". The EXTRACT function just works as expected, it extracts the requested nodes, but the result is still an XML fragment (a scalar value).
In order to achieve your goal, you have to break the MolTimeSeries sequence into relational rows.
Two similar solutions are possible, XMLTable (10gR2 and up) or Table/XMLSequence.
In your other post you mentioned db version 10.1, so I guess we'll go with XMLSequence :
SELECT 'RTE'
       ,to_date(substr(extractvalue(x.xml_data,'/MolDocument/ValidTimeInterval/@v'),1,16),'yyyy-mm-dd"T"hh24:mi')
       ,decode(extractvalue(x2.column_value,'/MolTimeSeries/Direction/@v'),'A01','Up','Down')
       ,to_number(regexp_substr(extractvalue(x2.column_value,'/MolTimeSeries/ContractIdentification/@v'),'\d+$'))
       ,to_number(extractvalue(x2.column_value,'/MolTimeSeries/Period/Interval/EnergyPrice/@v'))
       ,sysdate
       ,null
       ,null
       ,extractvalue(x2.column_value,'/MolTimeSeries/ContractIdentification/@v')
       ,'SUBMITTED'
       ,'A96'
       ,null
       ,sysdate
FROM xml_demo x,
     table(
       xmlsequence(
         extract(x.xml_data, '/MolDocument/MolTimeSeries')
     ) x2
WHERE existsnode(x.xml_data,'/MolDocument/DocumentIdentification[@v="MOL_20100331_1500_1600"]') = 1;Please also note the use of REGEXP_SUBSTR instead of the regular SUBSTR because it didn't work for "NG_20100331_1500_6".
Hope that helps.
Edited by: odie_63 on 24 juin 2010 21:18 - added regexp comment