Populate Parent/Child records with "Family" value

Hi,
I typically do the task I'm asking assistance with in Excel and it works great... the only problem is that it's VERY slow and takes a lot of work when over 1 million rows as you must split it up and yeah just painful! But what I'm trying to accomplish is
pretty straight forward for Excel, however escapes me in T-SQL! I have an ordered number that is sorted asc. I have a flag that indicates a record is the "Parent". I want to update the family value for the "Parent" and all sequential records
until the next "Parent" records is reached. So here is an example in Excel.
Using the following formula in cell C2 and pasting all the way to the last record, I can easily(for smaller sets anyway) create my desired results.
=IF(B2="Y",A2,C1)
Number Flag Family
1 Y 1
2 1
3 1
4 1
5 1
6 1
7 1
8 1
9 Y 9
10 9
11 Y 11
12 11
13 11
14 Y 14
15 14
16 14
17 14
18 14
19 14

I have an article on this exact topic
Fixing
Missing Data Based on Prior Row Information
It shows the concept of using OUTER APPLY to achieve desired functionality.
For every expert, there is an equal and opposite expert. - Becker's Law
My blog
My TechNet articles
I wish I knew how to apply this OUTER APPLY solution to an update statement so I could use. Would this be more efficient and faster than the solution provided by RyanAB? I have millions of records and NOT enough time... the fastest solution is definitely desired!!!
I ended up using this as a solution as it was the fastest. I first did:
SELECT a.Number, M.Number 'Family' into FamilyTable FROM Table a
OUTER APPLY (
SELECT TOP (1) *
FROM Table b
WHERE a.Number >= b.Number and b.[Flag] = 'Y'
ORDER BY b.[Number] DESC) M
ORDER BY 1, 2
Then updated the original like so:
update a set a.Family = b.Family
FROM Table a inner join FamilyTable b
on a.Number = b.Number
Thanks to all!!!

Similar Messages

Overriding VO Impl class methods to populate parent-child VO

Hi All,
I need to populate parent child VOs by giving webservice calls from respective impl classes.
Scenario : Lets say suppose i have 2 VOs as DepartmentVO and Employee VO with a viewLink on departmentId. Taking a particular instance, in Department table there are 2 rows and for each parent row we have 2 child rows in Employee.
Webservice Structure : I have a parent Object and inside that parent object i have list of child objects
e.g : public class DepartmentDTO{
private List<EmployeeDTO> employeeList;
After that i have extended Impl classes for both VOs and overriden method i.e. "executeQueryForCollection" to give a webservice call and get back response of DepartmentDTO[] with List<EmployeeDTO> inside it.
In DepartmenntVOImpl class i have overriden "createRowFromResultSet" which populates the parent VO with the number of parent records returned from webservice.
Code for Department impl class is :
protected ViewRowImpl createRowFromResultSet(Object qc,
ResultSet resultSet) {
Iterator deptIter= getPackageClassIterator(qc);
ViewRowImpl r = createNewRowForCollection(qc);
DepartmentDTO curDTO = (DepartmentDTO )deptIter.next();
populateAttributeForRow(r,
findAttributeDef("DeptId").getIndex(),
curDTO.geDeptId() );
populateAttributeForRow(r,
findAttributeDef("DeptName").getIndex(),
curDTO.getDeptName() );
return r;
Similarly i have overridden "createRowFromResultSet" method in EmployeeVOImpl to populate the child VO(this need to be populated from the list object inside parent object reponse)
Code as follows :
protected ViewRowImpl createRowFromResultSet(Object qc,
ResultSet resultSet) {
Iterator empItr= getPackageClassIterator(qc);
ViewRowImpl r = createNewRowForCollection(qc);
DepartmentDTO curDTO1 = (DepartmentDTO )empItr.next();
List<EmployeeDTO> list = curDTO1.getEmployeesList();
for (EmployeeDTO curDTO : list) {
populateAttributeForRow(r,
findAttributeDef("EmpId").getIndex(),
curDTO.getEmpId());
populateAttributeForRow(r,
findAttributeDef("EmpName").getIndex(),
curDTO.getEmpName() );
populateAttributeForRow(r,
findAttributeDef("EmpLastName").getIndex(),
curDTO.getEmpLastName() );
populateAttributeForRow(r,
findAttributeDef("DeptId").getIndex(),
curDTO.getDeptId());
return r;
Though the code looks simple and correct but the child records are not getting populated correctly. Am not sure if the code implementation to populate parent child VOs from webservcie is correct or not.
Would appreciate for Help or valuable Suggestions.
-Sanjeeb

Hi Ray,
Go for custom BADIs in this case. In fact, that is exactly the functionality you are looking for !
Just go through this excellent five part blog on Enhancement framework by Thomas Weiss.
What the New Enhancement Framework Is For Its Basic Structure and Elements For Beginners
The new Enhancement Framework Part 2 - What Else You Need to Know Before Building an Enhancement
How To Define a New BAdI Within the Enhancement Framework - Part 3 of the Series
How to implement a BAdI And How to Use a Filter - Part 4 of the Series on the New Enhancement Framework
Source Code Enhancements - Part 5 of the Series on the New Enhancement Framework
Cheers,
Mahesh

Script populate parent-child hierarchy not running

Hallo guys
I have problem when running script populate Parent-Child Relationship Table.
Here is the script.I don't even change the script.just run it in PL/SQL
declare
v_max_depth integer;
v_stmt varchar2(32000);
i integer;
begin
select max(level) into v_max_depth
from V_D_BRANCH
connect by prior BRANCH_KEY=PARENT_BRANCH_KEY
start with PARENT_BRANCH_KEY is null;
v_stmt := 'insert into DM_ANALYTICS.BRANCH_HIERARCHY (MEMBER_KEY, ANCESTOR_KEY, DISTANCE, IS_LEAF)
            select BRANCH_KEY as member_key, null, null, 0 from V_D_BRANCH where PARENT_BRANCH_KEY is null
            union all
            select member_key,
            replace(replace(ancestor_key,''\p'', ''|''), ''\'', ''\'') as ancestor_key,
            case when depth is null then 0
            else max(depth) over (partition by member_key) - depth + 1
            end as distance, is_leaf
from
select member_key,depth,
case     when depth is null then '' || member_key
when instr(hier_path, ''|'', 1, depth + 1) = 0 then null
   else substr(hier_path, instr(hier_path, ''|'', 1, depth) + 1, instr(hier_path, ''|'', 1, depth + 1) - instr(hier_path, ''|'', 1, depth) - 1)
end ancestor_key,
is_leaf
from
    (    select BRANCH_KEY as member_key, PARENT_BRANCH_KEY as ancestor_key, sys_connect_by_path(replace(replace(BRANCH_KEY, ''\'', ''\''), ''|'', ''\p''), ''|'') as hier_path,
      case when BRANCH_KEY in (select PARENT_BRANCH_KEY from V_D_BRANCH ) then 0 else 1 end as IS_LEAF
from V_D_BRANCH
connect by prior BRANCH_KEY = PARENT_BRANCH_KEY
start with PARENT_BRANCH_KEY is null
( select null as depth from dual;
for i in 1..v_max_depth - 1 loop
v_stmt := v_stmt || union all select '' || i || '' from dual;
end loop;
v_stmt := v_stmt || ) )
where ancestor_key is not null;
execute immediate v_stmt;
end;
but I got errors like following:
Error report -
ORA-06550: line 19, column 12:
PLS-00103: Encountered the symbol "insert into DM_ANALYTICS.BRANCH_HIERARCHY (MEMBER_KEY, ANCESTOR_" when expecting one of the following:
   ( - + case mod new not null <an identifier>
   <a double-quoted delimited-identifier> <a bind variable>
   continue avg count current exists max min prior sql stddev
   sum variance execute forall merge time timestamp interval
   date <a string literal with character set specification>
   <a number> <a single-quoted SQL string> pipe
   <an alternatively-quoted string
06550. 00000 - "line %s, column %s:\n%s"
*Cause:    Usually a PL/SQL compilation error.
*Action:
what should I do guys?any idea,any solution would be appreciated.thanks alot guys

956850 wrote:
hi,
i tried making "Dimension with Parent-Child Hierarchy" as described in your tutorials.
i finished with the admin tool and database changes and with no errors.
when trying to create a new answer, by only puting the hierarchy column in the answer.
the result show "no result" exist.
i am not sure what i am doing wrong. maybe you can point me to the problem?
thanks
Mirit.Hi Mirit,
What is the table that you created hierarchy on?
Which column in the hierarchy you pulled into your reports to see the No Result message?
Please query the table and see if has data.
Thanks,
G. SK

Parent/child records from same table

I want to create a query that is a union such that the 2nd resultset is based on the 1st resultset. I have a table that has parent/child records in the same table.
Table: EVENTS
EVENT_ID
PARENT_EVENT_ID
CREATED_DATE
(other columns)
if PARENT_EVENT_ID is null then it is a parent record, else it is a child record. I want to select all parent records then union them with all the associated child records...something like this:
select * from EVENTS where CREATED_DATE < sysdate - 90 and PARENT_EVENT_ID is null -- All parents
union
select * from EVENTS where PARENT_EVENT_ID in (select EVENT_ID from EVENTS where CREATED_DATE < sysdate - 90 and PARENT_EVENT_ID is null) -- include any children of parents selected from above
This works but it's kind of ugly, I want to avoid using the sub-select in the 2nd because it is a repeat of the 1st statement, is there a way to alias the first statement and just refer to it in the 2nd query?

Hi,
kev374 wrote:
Thanks, one question...
I did a test and it seems the child rows have to also satisfy the parent row's where clause, take this example:
EVENT_ID|PARENT_EVENT_ID|CREATED_DATE
2438 | (null) | April 9 2013
2439 | 2438 | April 11 2013
2440 | 2438 | April 11 2013
select * from EVENTS where CREATED_DATE < sysdate - 9
start with EVENT_ID = 2438
connect by PARENT_EVENT_ID = prior EVENT_IDSo you've changed the condition about only wanting roots and their children, and now you want descendants at all levels.
This pulls in record #2438 (per the sysdate - 9 condition) but 2439 and 2440 are not connected. Is there a way to supress the where clause evaluation for the child records? I just want to pull ALL child records associated with the parent and only want to do the date check on the parent.Since the roots (the only rows you want to exclude) have LEVEL=1, you can get the results you requested like this:
WHERE   created_date < SYSDATE - 9
OR      LEVEL         > 1However, since you're not ruling out the grandchildren and great-grandchildren any more, why wouldn't you just say:
SELECT *
FROM    events
WHERE   created_date     < SYSDATE - 9
OR      parent_event_id IS NOT NULL;?
CONNECT BY is slow. Don't use it if you don't need it.
If you x-reference my original query:
select * from EVENTS where CREATED_DATE < sysdate - 90 and PARENT_EVENT_ID is null -- All parents
union
select * from EVENTS where PARENT_EVENT_ID in (select EVENT_ID from EVENTS where CREATED_DATE < sysdate - 90 and PARENT_EVENT_ID is null) -- include any children of parents selected from above
The 2nd select does not apply the created_date < sysdate - 90 on the children but rather pulls in all related children :)Sorry; my mistake. That's what happens when you don't post sample data, and desired results; people can't test their solutions and find mistakes like that.

How to check/count whether child records with specific type exist?

Hi
We need to create a BIP report based on data from Siebel.
In Siebel we have two entities:
Entity <Mandate> = Parent Business Componet
Entity <Attachment> = Child Business Component
(An attachment has a specific type attribute (e.g. “contract”, “appendix” and lot of others…)
Relationship between Mandate and Attachments is 1-m.
We need to create a BIP report which displays all Mandates records which do not have at least two Attachments child records (there must be one attachment child record with type=”contract’ and another with type “appendix”). How can we check in BIP whether these child records with a specific Attachment type exist? And if not, display the Mandate in a list....
Many thanks
Alen

Thanks for your help
Well I'm not sure about the syatax I have to use in order to filter on the field *<GAMDocumentType>*
The structure of the xml we use is as follows:
<ListOfBipJbAmlMandateAttachmentReport>
<JbAmlMandate>
<ListOfJbAmlAttachment>
<JbAmlAttachment>
               <AccntFileName>DocumentPage_713328</AccntFileName>
               <GAMDocumentType>*contract</GAMDocumentType>
               <MandateId>1-4C79B</MandateId>
               <Status>Active</Status>
</JbAmlAttachment>
I tried <?for-each:JbAmlMandate[count(./JbAmlAttachment[GAMDocumentType='contract']) > 1 and count(./Attachment[type='appendix']) > 1]?>
But this returns nothing.
Many thanks for your help
Saggittarius

Union on tables with parent-child records and Sorting

Hi,
I have an application that has an existing query which returns org units (parent and child) from organization table with a sort on createddate + org_id combination
WITH Org_TREE AS (
SELECT *, null as 'IS_DELETED', convert (varchar(4000), convert(varchar(30),CREATED_DT,126) + Org_Id) theorderby
FROM Organization WHERE PARENT_Org_ID IS NULL and case_ID='43333'
UNION ALL
SELECT a1.*, null as 'IS_DELETED', convert (varchar(4000), a2.theorderby + convert(varchar(30),a1.CREATED_DT,126) + a1.Org_Id)
FROM Organization a1 INNER JOIN Org_TREE a2 ON a1.PARENT_Org_ID = a2.Org_Id and case_ID='43333'
SELECT * FROM Org_TREE order by theorderby
I have created a new log table for organization 'Organization_Log' with exact columns as Organization table with an additional 'IS_DELETED' bool column.
Questions:
I need to modiy the query,
1. To display the parent and child records both from the organization table and organization_log table.
2. the sort on the result should be based on 'Organization Name' column asc. First with parent org and the child org underneath it. For eg.
aaa
==>fff
==>ggg
bbb
==> aaa
==> hhh
Any help on how the query should be constructed?
Thanks
gkol

@Visakh16,
I am getting...
All queries combined using a UNION, INTERSECT or EXCEPT operator must have an equal number of expressions in their target lists.
The problem is that you will have different number of columns in both log and Organization and Organizationlog tables. UNION/UNION ALL expect the same number of columns.
Try the below:
WITH Org_Log_TREE AS (
SELECT Organization_name,Org_id,Parent_Org_id, IS_DELETED, CAST(Organization_Name AS varchar(max)) AS theorderby,1 AS level
FROM Organization_Log WHERE PARENT_Org_ID IS NULL and case_ID='43333'
UNION ALL
SELECT a1.Organization_name,a1.Org_id,a1.Parent_Org_id, a1.IS_DELETED, CAST(a2.theorderby +'/' + CAST(a1.Organization_Name AS varchar(1000)) AS varchar(max)),a2.Level + 1
FROM Organization_Log a1 INNER JOIN Org_Log_TREE a2 ON a1.PARENT_Org_ID = a2.Org_Id and case_ID='43333'
) ,Org_TREE AS (
SELECT Organization_name,Org_id,Parent_Org_id, NULL AS IS_DELETED, CAST(Organization_Name AS varchar(max)) AS theorderby,1 AS level
FROM Organization WHERE PARENT_Org_ID IS NULL and case_ID='43333'
UNION ALL
SELECT a1.Organization_name,a1.Org_id,a1.Parent_Org_id,NULL AS IS_DELETED, CAST(a2.theorderby +'/' + CAST(a1.Organization_Name AS varchar(1000)) AS varchar(max)),a2.Level + 1
FROM Organization a1 INNER JOIN Org_TREE a2 ON a1.PARENT_Org_ID = a2.Org_Id and case_ID='43333'
SELECT * FROM Org_Log_TREE
UNION ALL
SELECT * FROM Org_TREE
ORDER BY LEFT(theorderby,CHARINDEX('/',theorderby + '/')-1),Level

How to build what a parent child record looked like from audit history

I've got a simple parent child relationship (eg DEPT/EMP) and I have audit triggers on the 2 tables that log any changes. The 2 tables are seen by the User as 1 entity, so if an EMP record changes, that is essentially a DEPT change. I want to show a record where the info shown is correct at that particular point in time.
Now for a very crude example to illustrate...
So, says it's 9:00 and we have:
Dept 1=Sales
Emp 1=Clark
Emp 2=Smith
Now we change Emp 1 at 9:01 to Smythe
Next we change Dept 2 at 9:02 to Sales UK
Next we add Emp 3=Jones at 9:04
Next we change Emp1 at 9:05 to Clarke
So now if the User looks at the history of Dept 1 they will see it has changed at the following times:
9:01 (emp change)
9:02 (dept change)
9:04 (emp change)
9:05 (emp change)
even though the actual Dept table was only changed once.
This part is easy enough to extract but the next part I'm struggling to get my head around (must be Friday afternoon!).
The User wants to be able to view what the "Dept" looked like at the time of the change. So we would get the following results:
9:00 Dept 1=Sales, Emp 1=Clark, Emp 2=Smith (as per original data)
9:01 Dept 1=Sales, Emp 1=Clark, Emp 2=Smythe
9:02 Dept 1=Sales UK, Emp 1=Clark, Emp 2=Smythe
9:04 Dept 1=Sales UK, Emp 1=Clark, Emp 2=Smythe, Emp 3=Jones
9:05 Dept 1=Sales UK, Emp 1=Clarke, Emp 2=Smythe, Emp 3=Jones
I'm trying to write a bit of SQL to do this but the tricky bit comes in due to having 2 different tables that have triggered the change to a "DEPT". I see the output as being as many rows as there are children, so the raw output for the above summary would be along the lines of:
9:00, Dept 1=Sales, Emp 1=Clark
9:00, Dept 1=Sales, Emp 2=Smith
9:01, Dept 1=Sales, Emp 1=Clark
9:01, Dept 1=Sales, Emp 2=Smythe
9:02, Dept 1=Sales UK, Emp 1=Clark
9:02, Dept 1=Sales UK, Emp 2=Smythe
9:04, Dept 1=Sales UK, Emp 1=Clark
9:04, Dept 1=Sales UK, Emp 2=Smythe
9:04, Dept 1=Sales UK, Emp 3=Jones
9:05, Dept 1=Sales UK, Emp 1=Clarke
9:05, Dept 1=Sales UK, Emp 2=Smythe
9:05, Dept 1=Sales UK, Emp 3=Jones
Any help appreciated!!
An alternative to all this is to make my trigger on each table populate this info (ie write whatever is currently in place) into a table structured as I have mentioned above (so Dept i denormalised) but that looks to be a bit messy.

Two tables, Dept and Emp. Both have audit triggers inserting into it's own audit table.
create or replace table DEPT(
dept_no number,
dept_name varchar2(50),
dept_city varchar2(50));
create or replace table EMP(
emp_no number,
emp_surname varchar2(100),
emp_firstname varchar2(50),
emp_dept_no number);dept_no pk of DEPT
emp_no pk of EMP
emp_dept_no FK from EMP to DEPT.dept_no
create or replace table DEPT_AUDIT(
dept_no number,
dept_name varchar2(50),
dept_city varchar2(50),
audit_date_time date);
create or replace table EMP_AUDIT(
emp_no number,
emp_surname varchar2(100),
emp_firstname varchar2(50),
emp_dept_no number,
audit_date_time date)
create or replace trigger dept_audit
after insert or update
on dept for each row
begin
insert into dept_audit(dept_no, dept_name, dept_city, audit_date_time)
values(:NEW.dept_no, :NEW.dept_name, :NEW.dept_city, sysdate);
end;
create or replace trigger emp_audit
after insert or update
on emp for each row
begin
insert into emp_audit(emp_no, emp_surname, emp_firstname, emp_dept_no, audit_date_time)
values(:NEW.emp_no, :NEW.emp_surname, :NEW.emp_firstname, :NEW.emp_dept_no, sysdate);
end;My above example (ignoring that the columns don't match up but hopefully you get the drift) would end up with the following results in the 2 audit tables:
DEPT_AUDIT
9:00 Dept 1=Sales (insert)
9:02 Dept 1=Sales UK (update)
EMP_AUDIT
9:00 Emp 1=Clark (insert)
9:00 Emp 2=Smith (insert)
9:01 Emp 2=Smythe (update)
9:04 Emp 3=Jones (insert)
9:05 Emp 1=Clarke (update)
As I said originally, I want to get a view of what the "DEPT" looked like at a particular time and to do this I need to take into account the EMP changes for the DEPT which is why my results would be as follows:
9:00 Dept 1=Sales, Emp 1=Clark, Emp 2=Smith
9:01 Dept 1=Sales, Emp 1=Clark, Emp 2=Smythe
9:02 Dept 1=Sales UK, Emp 1=Clark, Emp 2=Smythe
9:04 Dept 1=Sales UK, Emp 1=Clark, Emp 2=Smythe, Emp 3=Jones
9:05 Dept 1=Sales UK, Emp 1=Clarke, Emp 2=Smythe, Emp 3=Jones
Though what I really want is the rows that make it up like:
9:00, Dept 1=Sales, Emp 1=Clark
9:00, Dept 1=Sales, Emp 2=Smith
9:01, Dept 1=Sales, Emp 1=Clark
9:01, Dept 1=Sales, Emp 2=Smythe
9:02, Dept 1=Sales UK, Emp 1=Clark
9:02, Dept 1=Sales UK, Emp 2=Smythe
9:04, Dept 1=Sales UK, Emp 1=Clark
9:04, Dept 1=Sales UK, Emp 2=Smythe
9:04, Dept 1=Sales UK, Emp 3=Jones
9:05, Dept 1=Sales UK, Emp 1=Clarke
9:05, Dept 1=Sales UK, Emp 2=Smythe
9:05, Dept 1=Sales UK, Emp 3=Jones
Hope that is a bit clearer.
thanks for taking time to look.

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

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>

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

How to achieve parent - child record insert in forms

Guys,
I am new to forms. I do have a requirement where I need to create one tabular block for 10 rows (parent record) and then I need to create three tabular tabs of 5 rows for detailed records (child). That means each tab can have 5 records for 1 parent row.. Thats how this parent-child structure has ONE to MANY relationships
To achieve this requirement I created one parent block of table type (using table XX_Parent) and then I created three detailed block for each tab (Tab_A,Tab_B,Tab_C) using child table (XX_Child). For each tab while inserting the data I took one context_value column in child table. i.e. when I insert the data using Tab_A , context_value column will be A. In the same way B and C for Tab_B for Tab_C respectively. And I wrote this logic at PRE-INSERT trigger for each tab. But when I insert the data for all the tabs, it allows me to enter successfully for Tab_A only, the moment I navigate from Tab_A to enter the data intoTab_B, it throws me error like "You are passing NULL value to set context_value". I hope I make the requirement clear to you.
I am not sure if I am following the right approach to achieve my requirement. if its not right, please suggest me right one.. and if its right please guide me resolve the error..
Looking forward to your reply.
Thanks
Sunil

Andreas,
I had already created relations between Matster Block -Tab_A block, Master Block -Tab_B block and Master Block - Tab_C block...
The point is all the three tabs are created using one table XX_Child only.. I used context_value A,B and C respectively so that once I query the data for XX_Child from backend, I should be in a position to figure out what data is inserted using Tab A, B and C. why I need so .. because If take my entire form in query mode and search for a master record.. then Tab_A, tab_B and Tab_C display their respective data. I hope, I make you understand.. Can you please guide me.
Thanks
Sunil

Security on Employee Parent Child Dimension with SCD 2

Hi
I am looking for a solution for how to get all child members of a Parent employee including the one in the history. I tried multiple ways the problem which I am facing are
Scenario 1: When I am using this approach, since the Employee with 'Jump' as employee id do have a history record with no children under him I am getting 'Null'. But there are new records (last 3 SCD changes) where he has employees under him.
SELECT descendants(
// Check if current employee is a manager
iif(hierarchize(exists([Employee].[Manager Employee Dim ID].members,
(strtomember("[Employee].[Employee Login ID].&[SIC\jump]"))),post).item(0).item(0).parent.datamember is
hierarchize(exists([Employee].[Manager Employee Dim ID].members,
(strtomember("[Employee].[Employee Login ID].&[SIC\jump]"))), post).item(0).item(0),
// End of check if current employee is manager
hierarchize(exists([Employee].[Manager Employee Dim ID].members,
(strtomember("[Employee].[Employee Login ID].&[SIC\jump]"))),post).item(0).item(0).parent,
hierarchize(exists([Employee].[Manager Employee Dim ID].members,
(strtomember("[Employee].[Employee Login ID].&[SIC\jump]"))), post).item(0).item(0))) ON 0,
[Measures].[Aggregate Count] On 1
FROM [Data Ware House]
Scenario 2: When I put a filter (as a set) to check if the record is current then I am getting the latest record for employee 'Jump'. But here the problem is that the employee under him I am only getting the [Measures].[Aggregate Count] of the latest
change only.
SELECT descendants(
// Check if current employee is a manager
iif(hierarchize(exists([UW Employee].[Manager Employee Dim ID].members,
(strtomember("[UW Employee].[Employee Login ID].&[SIC\flew]"),[UW Employee].[Current Flag].&[Y])),
post).item(0).item(0).parent.datamember is
hierarchize(exists([UW Employee].[Manager Employee Dim ID].members,
(strtomember("[UW Employee].[Employee Login ID].&[SIC\flew]"),[UW Employee].[Current Flag].&[Y])), post).item(0).item(0),
// End of check if current employee is manager
hierarchize(exists([UW Employee].[Manager Employee Dim ID].members,
(strtomember("[UW Employee].[Employee Login ID].&[SIC\flew]"),[UW Employee].[Current Flag].&[Y])),
post).item(0).item(0).parent,
hierarchize(exists([UW Employee].[Manager Employee Dim ID].members,
(strtomember("[UW Employee].[Employee Login ID].&[SIC\flew]"),[UW Employee].[Current Flag].&[Y])), post).item(0).item(0))
) ON 0,
[Measures].[Contract Change Aggregate Count] ON 1
FROM [Data Ware House]
If any one can help me out with this that will be helpful.
What I need is all the members under the employee 'Jump', including the Old scd records for 'jump'
regards
Robin
Robs

Hi,
Outline is blank at the start of load and I expect dimensions Product and ProdAttribute to be added.
Generations are unapplicable, since I can't turn this parent-child dimension (it's ragged) into generation one.
And, by the way Integration Services manages to do this somehow, why rule files wouldn't?

OBIEE Parent-Child Dimension with multiple hierarchies

Is it possible to have in a parent-child dimension multiple hierarchies? I want to have the same botom childs to aggregate to different parrents. example.
Account Dimension
HR1-total
HR1-parrent1
CHILD1
HR1-parrent2
CHILD2
HR2-total
HR2-parrent1
CHILD1
CHILD2
Thanks in advance.

No, it is not possible to have multiple hiearchies in a PC logical dimension.
The concept of the Parent-Child (PC) hierarchy is completely different from the level-based hierarchy. Specifically the PC hierarchy expects a predefined / architected table with corresponding PC column/value structuring with or without attributes.
Short story even shorter it is not possible.
Longer...
In the RPD the BMM actually prevents you from adding a new logical level, child level, or parent level when you have selected that the logical dimension be a parent child logical dimension with Parent-Child hierarchy.
On another note...
Have you tried architecting/building your PC source table so that it represents the roll-up and bottom child-levels like you are seeking? That is, that same member in the child column more than once having a separate parent value. That should be doable.
That was a great question, please award points if this answered your question or it was helpful.
Cheers,
Christian
http://www.artofbi.com

Cdc- deletion of parent- child records

Hi,
I am working with CDC-consistent feature in odi.
Here my scenario is, I have a record say 120 (primary key) in table A(parent source table) and it is used as a foreign key in Table B.
both child and parent are inserted into the concerned tables of target.
Now i want to delete this 120 record from target parent and child tables.
IN the pkg i arranged the pkg scenarios as follows
odiwaitforlogdata----->source model(with extenwindow and lock subscriber option selected)------>parent pkg scenario------>child pkg scenario----->source mode(with unlock subscriber and purge journal options). ------------> This works fine for insert and update.
odiwaitforlogdata----->source model(with extenwindow and lock subscriber option selected)------>child pkg scenario--------->parent pkg scenario----->source mode(with unlock subscriber and purge journal options). ------------> This works fine for delete.
Can't I achieve these two in one pkg
Please Guide.
Regards,
Chaitanya.

Hi,
kev374 wrote:
Thanks, one question...
I did a test and it seems the child rows have to also satisfy the parent row's where clause, take this example:
EVENT_ID|PARENT_EVENT_ID|CREATED_DATE
2438 | (null) | April 9 2013
2439 | 2438 | April 11 2013
2440 | 2438 | April 11 2013
select * from EVENTS where CREATED_DATE < sysdate - 9
start with EVENT_ID = 2438
connect by PARENT_EVENT_ID = prior EVENT_IDSo you've changed the condition about only wanting roots and their children, and now you want descendants at all levels.
This pulls in record #2438 (per the sysdate - 9 condition) but 2439 and 2440 are not connected. Is there a way to supress the where clause evaluation for the child records? I just want to pull ALL child records associated with the parent and only want to do the date check on the parent.Since the roots (the only rows you want to exclude) have LEVEL=1, you can get the results you requested like this:
WHERE   created_date < SYSDATE - 9
OR      LEVEL         > 1However, since you're not ruling out the grandchildren and great-grandchildren any more, why wouldn't you just say:
SELECT *
FROM    events
WHERE   created_date     < SYSDATE - 9
OR      parent_event_id IS NOT NULL;?
CONNECT BY is slow. Don't use it if you don't need it.
If you x-reference my original query:
select * from EVENTS where CREATED_DATE < sysdate - 90 and PARENT_EVENT_ID is null -- All parents
union
select * from EVENTS where PARENT_EVENT_ID in (select EVENT_ID from EVENTS where CREATED_DATE < sysdate - 90 and PARENT_EVENT_ID is null) -- include any children of parents selected from above
The 2nd select does not apply the created_date < sysdate - 90 on the children but rather pulls in all related children :)Sorry; my mistake. That's what happens when you don't post sample data, and desired results; people can't test their solutions and find mistakes like that.

Load Parent-Child dimension with attributes by rule file

Well, I've cracked my head open while banging against this wall.
Task is plain & simple:
How can I load build dimension from such table, using rule file?
Data
ProductParent ProductChild ProductAttribute
Balls Ball_8 Round
I want to load both product dimension and attribute by single rule file (neither of these dimensions exists in database before load), so I do the following:
1) Set dimension build settings, creating dimension Products (marking it P\C generated, sparse), having an attribute dimension AttributeDim
2) Mark the fields as Parent, Child and AttributeDim of dimension Products
After loading rule file a get three types of errors:
a) "Base member (Balls) association level does not match base dimension association level" -- does this mean that I cannot assign attributes in P\C way?
b) "Record #327 - Can not uniquely identify member by Balls (3335)
Balls Ball_8 Round" -- That's strange, I thought member should be identified by Ball_8 (and this column is unique)
c) "\\Record #418 - Error in association transaction [Ball_8] to [Round] (3362)"
I've read all I could find on topic from this forum, network45 board and going to start my way through metalink3, but maybe some kind soul will rescue me from drowning in information ocean?

Hi,
Outline is blank at the start of load and I expect dimensions Product and ProdAttribute to be added.
Generations are unapplicable, since I can't turn this parent-child dimension (it's ragged) into generation one.
And, by the way Integration Services manages to do this somehow, why rule files wouldn't?

How to model parent child relationship with DPL? @Transient?

Hello All,
I want to model a parent entity object with a collection of child entities:
@Entity
public class Parent{
@PrimaryKey
String uuid;
List<Child> children;
@Entity
public class Child{
@PrimaryKey
String id;
I know that the DPL won't support automatic persistence where it'll recursively go through my parent bean and persist my children with one call. Is there a way of applying the equivalent to JPA's @Transient annotation on "children" so I can persist the children manually and have the engine ignore the collection?
If not and I want to return to the user a Parent with a List named "children," do I have to create a new object which is identical to Parent, but doesn't have the BDB annotations and manually assemble everything? If possible, I'd like to avoid defining redundant objects.
Thanks in advance,
Steven
Harvard Children's Hospital Informatics Program
Edited by: JavaGeek_Boston on Oct 29, 2008 2:22 PM

Hi Steven,
The definition of persistence is here:
http://www.oracle.com/technology/documentation/berkeley-db/je/java/com/sleepycat/persist/model/Entity.html
And includes this: "All non-transient instance fields of an entity class, as well as its superclasses and subclasses, are persistent. static and transient fields are not persistent."
So you can use the Java transient keyword. If that isn't practical because you're using transient in a different way for Java serialization, see the JE @NotPersistent annotation.
In general a parent-child relationship between entities is implemented almost as you've described, but with a parentId secondary key in the Child to index all children by their parent. This enables a fast lookup of children by their parent ID.
I suggest looking at this javadoc:
http://www.oracle.com/technology/documentation/berkeley-db/je/java/com/sleepycat/persist/SecondaryIndex.html
as it describes all types of entity relationships and the trade-offs involved. The department-employee relationship in these examples is a parent-child relationship.
--mark

Hide parents in a parent-child hierarchy with dynamic rights

Hi all,
I'm facing the following issue with dynamic rights on a parent-child hierarchy:
scenario:
- a cube with a company parent-child hierarcy
- dynamic security on through a role on dimension data based on a list with user (login names) and their
departments, with Enable
Visual Total is switched ON
- Multiple facts connected to this hierarchy.
When a user connects to the cube (in this case with Excel or through a dashboard in Performance Point services) he/she is allowed to see the facts belonging to theirs departments and all child departments, so far so good. But the totals of the facts are
repeated for all higher parents. Now i know thats nog 'wrong' but our users think thats very confusing in some cases, so is there a way
- to hide the parents in the hierarhcy
- or to set all the parent measures to null/empty?
I prefer to make a solution in the cube so that all client programs can benefit from this solution.
thanx in advance for all input!

so that User A have scope: [Dimension_Name].[Attribute XYZ] and User B: [Dimension_Name].[Attribute ABC] ?
Hi Wilfred,
If I understanding correctly, you want different user can see different dimension attributes based on the different right on role, right?
Based on my research, it seems that there is no a directly way to achieve this requirement. What we can do is that "Deselect all members" for each attribute of the dimension in Dimension Data tab, which will hide all members of the dimension for the role.
Even the dimension metadata is visible in the cube in clients, its members can't be seen. Here are some useful links for your reference.
http://saysmymind.wordpress.com/2013/01/15/hide-a-specific-attribute-hierarchy-in-role-playing-dimension-ssas-multidimensional/
http://social.msdn.microsoft.com/Forums/sqlserver/en-US/e2596eec-7c2d-48c0-8e81-538a8c632360/hide-dimension-based-on-roles
Hope this helps.
Regards,
Charlie Liao
TechNet Community Support

Populate Parent/Child records with "Family" value

Similar Messages

Maybe you are looking for