WRITING A GENERAL qUERY

Hi
I want to write a query which takes a argument such as a number n.
The query is used to get any nth largest number from a table.
I could write it for only one say getting the second largest number. I wrote it as:
A= Select * from Emp where sal !=(select max(sal) from Emp);
select max(sal) from A;
It is very complex to make this query a General one.
Can any body suggest an idea.
thanks

select ename, sal from (select row_number() over (order by sal desc) r, ename, sal from emp) where r=&n;
Enter value for n: 5
ENAME SAL
BLAKE 2850
Please read the documentation about the differences between rank(), dense_rank() and rownumber()
Regards
Laurent Schneider
OCM DBA

Similar Messages

General Query to identify parent columns and child columns in a table

Does anyone know a general query that can be run to identify the child records associated with the parent column within a single table.....and between other tables?
Am I correct in assuming the parent column is the 'primary key'?
Thanks.....I'm a new to oracle...and need some help understanding
my company's crazy DB structure

You can use
User_Constraints
User_Cons_Columns
views to identify parent and child table columns
SELECT * FROM User_Constraints WHERE Constraint_Type = 'R' AND Table_Name = '<TABLENAME>';
SELECT * FROM User_Cons_columns WHERE Constraint_Name = '<Name from Above query>';
will give you columns of the parent table (if you use value from constraint_name) and of child table (if you use value from r_constraint_name).
HTH..

General Query in OSB - 10gR3

Hi,
I have a general query in OSB. Whenever I try to add any expression through any message processing controls, there are three other tabs, comments, namespaces and variables. Can someone explain what is the use of namespaces and variables. If possible with the help of an example.
Thanks,
Bals

Hi Bals,
Comments allow the developer to put some lines regarding the functionality of that action/stage. This is same as putting comment in any programming language.
XML Namespaces provide a method to avoid element name conflicts in a XML. To know more please refer -
http://www.w3schools.com/xml/xml_namespaces.asp
Variables section refers to the message context variables like $header, $body, $inbound, $outbound etc. You may refer below link to know more -
http://download.oracle.com/docs/cd/E13159_01/osb/docs10gr3/userguide/context.html
Regards,
Anuj

Need help in writing SP notification Query

Hi,
My purpose of writing a query is to restrict a user to add journal entry in case account code lies in Expense Drawer and Cost Center Code is not in user define range.
Please help me on this:-
IF (@object_type = '30' AND (@transaction_type = 'A' Or @transaction_type = 'U'))
BEGIN
Declare @AccountCodeMaster as Nvarchar(20)
Declare @AccountonDocument as Nvarchar(20)
Declare @ProfitCentreMaster as Nvarchar(20)
Declare @ProfitCentreOnDocument as Nvarchar(20)
Set @AccountCodeMaster = (select AcctCode from oact where GroupMask = 5)
Set @AccountonDocument = (Select Account from jdt1 where TransId = @list_of_cols_val_tab_del)
Set @ProfitCentreMaster = (select PrcCode from ocr1 where PrcCode >= 91100001 and PrcCode <= 91921001)
Set @ProfitCentreOnDocument = (select ProfitCode from jdt1 where TransId = @list_of_cols_val_tab_del)
if (@AccountonDocument) in (@AccountCodeMaster) and (@ProfitCentreOnDocument) not in (@ProfitCentreMaster)
Begin
set @error =1
set @error_message = 'Profit Center is wrong,Please select the right profit Center else Consult to Finance Department'
End
END
Thanks

Hi,
Welcome you post on the forum.
You may try:
IF (@object_type = '30' AND @transaction_type in ('A', 'U')
BEGIN
If Exists (Select T0.Account from jdt1 T0 inner join oact T1 ON T1.AcctCode=T0.Account AND T1.GroupMask = 5
where T0.TransId = @list_of_cols_val_tab_del AND T0.ProfitCode NOT IN (select PrcCode from ocr1 where PrcCode >= 91100001 and PrcCode <= 91921001))
Begin
set @error =1
set @error_message = 'Profit Center is wrong,Please select the right profit Center else Consult to Finance Department'
End
END
Thanks,
Gordon

HOWTO: Writing Out XML Query Results of Any Size in Java with XML SQL Utility

A customer mailed me asking for an example of how to use our XML SQL Utility to write out query results for tons of query result rows.
With tons of data, the getXMLDOM() and getXMLString() methods are not really appropriate due to the size.
The XML SQL Utility offers a getXMLSAX() method that streams SAX2 events to report the data being queried. This is the approach we can use to handle data of any size.
It dawned on me today that by putting together the XML SQL Utilities getXMLSAX() routine, and the oracle.xml.parser.v2.XSLSAXPrintDriver SAX2 ContentHandler, we can effectively stream out data of any length to an appropriate writer.
Here's a code example to get the point across:
package test;
import java.io.BufferedOutputStream;
import java.io.PrintWriter;
import java.sql.Connection;
import java.sql.Driver;
import java.sql.DriverManager;
import java.sql.SQLException;
import java.util.Properties;
import javax.xml.transform.OutputKeys;
import oracle.jdbc.OracleDriver;
import oracle.xml.parser.v2.XSLException;
import oracle.xml.parser.v2.XSLOutput;
import oracle.xml.parser.v2.XSLSAXPrintDriver;
import oracle.xml.sql.query.OracleXMLQuery;
public class Example {
private static final String QUERY = "select * from emp";
public static void main(String[] args) throws Throwable {
      Connection conn = getConnection();
      OracleXMLQuery q = new OracleXMLQuery(getConnection(),QUERY);
      // Any printwriter will do. Here's we're output to standard out.
      PrintWriter output = new PrintWriter(new BufferedOutputStream(System.out));
      // This is a SAX2 Content Handler used by the Oracle XSLT Engine
      // to serialize a stream of sax2 events as an XML document
      // We'll use it to serialize the sax2 events from the XML SQL Utility
      // out as an XML document.
      XSLSAXPrintDriver ch = new XSLSAXPrintDriver(output, outputOptions());
      // This asks XML SQL Utility to fire sax events for the data
      // being fetched instead of creating DOM nodes or returning text.
      // By using the XSLSAXPrintDriver content handler, these events
      // get handled by writing them directly to the output stream
      q.getXMLSAX(ch);
      ch.flush();
      q.close();
      conn.close();
private static XSLOutput outputOptions() throws XSLException {
    XSLOutput x = new XSLOutput();
    Properties props = new Properties();
    props.put(OutputKeys.METHOD,"xml");
    props.put(OutputKeys.INDENT,"yes"); // Set to "no" for non-indented
    x.setProps(props);
    return x;
public static Connection getConnection() throws SQLException {
    String username = "scott";
    String password = "tiger";
    String thinConn = "jdbc:oracle:thin:@localhost:1521:ORCL";
    Driver d = new OracleDriver();
    return DriverManager.getConnection(thinConn,username,password);
}

Hi Uber,
This is a known issue that error occurs when running report "Count of instances of specific software registered with Add or Remove Programs" due to non-printable characters for XML. Based on internal research, the hotfix for this issue will be
included in the System Center 2012 Configuration Manager Service Pack 1.
As a workaround, you can remove the nonprintable character populated into the report parameter by referring to the following KB article:
http://support.microsoft.com/KB/914159
Hope this helps.
Regards,
Mike Yin
Mike Yin
TechNet Community Support

Frozen, white screen w/ black writing after general software update.

I was trying to do a general software update on my MackBook Air when halfway through, the update stopped and a white screen popped up with the Apple in the middle and some black writing appeared in the upper left-hand corner. The writing is nothing that I recognize and looks like some kind fo code. The few things that make sense are " Unable to find driver for this platform" "BSD process name corresponding to current thread: Unknown" and "System update time in nanoseconds 347370409." I cannot do ANYTHING with the computer except turn it on and off and when I do that, the same screen comes up each time I turn it on. I cannot click on anything because no cursor appears. Please help! Thank you!

Hi mary1490,
Thanks for using Apple Support Communities. I would recommend troubleshooting this as similar to a kernel panic. This article has some steps that may help:
OS X Mavericks: If your Mac restarts and a message appears
http://support.apple.com/kb/PH14063
Cheers,
- Ari

Syntax error while writing a select query.

Hi all,
I have a requirement where I have to pick a value if the text for that value is
MYCARu2019S Lovliest Car.
And so I wrote a query that
SELECT     RUECK INTO XRUECK FROM AFVC WHERE ltxa1 = ' MYCARu2019S Lovliest Car'.
                ENDSELECT.
But it gives me a syntax error saying
u201CLiterals taking up more than one line not permittedu201D.
Can some one tell me what is wrong. I need to select RUECK value from AFVC table if
ltxa1 value is ' MYCARu2019S Lovliest Car'.
Kindly help what is going wrong...
Regards,
Jessica Sam

Hi,
Narendran is right use two single quotes.
also careful wile comparing string. i think as you write ltxa1 value is ' MYCARu2019S Lovliest Car'.
so while comparing you must not use space just before staring use the following
SELECT rueck INTO xrueck FROM afvc WHERE ltxa1 = 'MYCAR''S Lovliest Car'. " Not use Space before M
ENDSELECT.
Hope will help you.
Kind Regards,
Faisal

Need help on performing database queries with a general query

Dear experts,
I have two issues
1)
I have developed a software which connects to Oracle database
performing and executing DDL/DML statements.
I am using a Jdbc thin oracle driver.
When i perform a wrong syntax query or say a query using invalid column.I get exception accordingly.However,I want to know if there is any way by which i can know the exact position where i am making error
just as an asterik(*) comes in sql plus in case of invalid column or so.
2)
Whenever i minimize software window (made of swing components)
and again maximize it.It takes almost 12 secs to be visible otherwise it remains white.Is swing that much slow ???

(1) No.
(2) It's possible to make Swing programs that slow. (I know, I have done it.) But usually the slowness is in application code (your code) and not in Swing itself.

Need help in writing a select query to pull required data from 3 tables.

Hi,
I have three tables EmpIDs,EmpRoles and LatestRoles. I need to write a select Query to get roles of all employees present in EmpIDs table by referring EmpRoles and LatestRoles.
The condition is first look into table EmpRoles and if it has more than one entry for a particular Employee ID than only need to get the Role from LatestRoles other wise consider
the role from EmpRoles .
Sample Script:
Create Table #EmpIDs
(EmplID int )
Create Table #EmpRoles
(EMPID int,Designation varchar(50))
Create Table #LatestRoles
EmpID int,
Designation varchar(50)
Insert into #EmpIDs values (1),(2),(3)
Insert into #EmpRoles values (1,'Role1'),(2,'Role1'),(2,'Role2'),(3,'Role1')
Insert into #LatestRoles values (2,'Role2')
Employee ID 2 is having two roles defined in EmpRoles so for EmpID 2 need to fetch Role from LatestRoles table and for
remaining ID's need to fetch from EmpRoles .
My Final Output of select query should be like below.
EmpID Role
1 Role1
2 Role2
3 Role1
Please help.
Mohan

Mohan,
Can you check if this answers your requirement:
Create Table #EmpIDs
(EmplID int )
Create Table #EmpRoles
(EMPID int,Designation varchar(50))
Create Table #LatestRoles
EmpID int,
Designation varchar(50)
Insert into #EmpIDs values (1)
Insert into #EmpIDs values (2)
Insert into #EmpIDs values (3)
Insert into #EmpRoles values (1,'Role1')
Insert into #EmpRoles values (2,'Role2')
Insert into #EmpRoles values (2,'Role1')
Insert into #EmpRoles values (3,'Role1')
Insert into #LatestRoles values (2,'Role2')
--Method 1
select e.EmplID,MIN(ISNULL(l.Designation,r.Designation)) as Designation
from #empids e
left join #emproles r on e.emplID=r.EmpID
left join #latestRoles l on e.emplID=l.EmpID
group by e.EmplID
--Method 2
;with cte
as
select distinct e.EmplID,r.Designation,count(*) over(partition by e.emplID) cnt
from #empids e
left join #emproles r on e.emplID=r.EmpID
select emplID,Designation
from cte
where cnt=1
UNION ALL
select a.EmplID,l.Designation
from
(select distinct EmplID from cte where cnt>1) a
join #Latestroles l on a.EmplID=l.EmpID
order by emplID
Thanks,
Jay
<If the post was helpful mark as 'Helpful' and if the post answered your query, mark as 'Answered'>

Problem writing a sql query for a select list based on a static LOV

Hi,
I have the following table...
VALIDATIONS
ID          Number     (PK)
APP_ID          Number
REQUESTED     Date
APPROVED     Date
VALID_TIL     Date
DEPT_ID          Number     (FK)
I have a search form with the following field item variables...
P11_DEPT_ID (select list based on dynamic LOV from depts table)
P11_VALID (select list based on static Yes/No LOV)
A report on the columns of the Validations table is shown based on the values in the search form. So far, my sql query for the report is...
SELECT v.APP_ID,
v.REQUESTED,
v.APPROVED,
v.VALID_TIL,
d.DEPT
FROM DEPTS d, VALIDATIONS v
WHERE d.DEPT_ID = v.DEPT_ID(+)
AND (d.DEPT_ID = :P11_DEPT_ID OR :P11_DEPT_ID = -1)
This query works so far. My problem is that I don't know how to do a search based on the P11_VALID item - if 'yes' is selected, then the VALID_TIL date is still valid. If 'no' is selected then the VALID_TIL date has passed.
Can anyone help me to extend my query to include this situation?
Thanks.

Hello !
Let's have a look at my example:create table test
id        number
,valid_til date
insert into test values( 1, sysdate-3 );
insert into test values( 2, sysdate-2 );
insert into test values( 3, sysdate-1 );
insert into test values( 4, sysdate );
insert into test values( 5, sysdate+1 );
insert into test values( 6, sysdate+2 );
commit;
select * from test;
def til=yes
select *
from test
where decode(sign(trunc(valid_til)-trunc(sysdate)),1,1,0,1,-1)
       =decode('&til','yes',1,-1);
def til=no
select *
from test
where decode(sign(trunc(valid_til)-trunc(sysdate)),1,1,0,1,-1)
       =decode('&til','yes',1,-1);
drop table test; It's working fine, I've tested it.
The above changes to my first idea I did because of time portion of the DATE datatype in Oracle and therefore the wrong result for today.
For understandings:
1.) TRUNC removes the time part of DATE
2.) The difference of to date-values is the number of days between.
3.) SIGN is the mathematical function and gives -1,0 or +1 according to an negative, zero or positiv argument.
4.) DECODE is like an IF.
Inspect your LOV for the returning values. According to my example they shoul be 'yes' and 'no'. If your values are different, you may have to modify the DECODE.
Good luck,
Heinz

Help needed on writing a SQL query

Here is my table that shows records of 3 ORDER_ID (10, 20 and 30). All I need to do is, pick the first record of each order_id and check the event_id, if the event_id is same for the next record, ignore it, else show it and ignore rest all records for that order_id. This way my query output will show only two records for each ORDER_ID. Query should produce records that are in BOLD in the sample data. (Database - 11g)
Thanks for your help in advance
ORDER_ID     EVENT_ID     EVNT_DATE     STATE_CODE
*10     16937555     20100212     COMPLETE*
10     16937555     20100212     ACTIVE
*10     16308004     20100129     OCCURRED*
10     16131904     20100125     ACTIVE
10     16270684     20100128     OCCURRED
10     14899116     20091213     ACTIVE
10     16085672     20100123     COMPLETE
10     16085673     20100123     OCCURRED
10     14899119     20100123     COMPLETE
10     14899120     20100123     COMPLETE
*20     17134164     20100223     COMPLETE*
20     17134164     20100223     ACTIVE
20     17134164     20100223     STARTED
*20     15479131     20100105     OCCURRED*
20     15478409     20100105     OCCURRED
20     15478408     20100105     ACTIVE
20     15119404     20100105     COMPLETE
20     14346123     20091129     ACTIVE
20     15467821     20100104     OCCURRED
20     14346125     20091216     COMPLETE
20     14346126     20091215     COMPLETE
20     14346126     20091214     COMPLETE
*30     18814670     20100412     COMPLETE*
30     18814670     20100412     ACTIVE
*30     18029509     20100320     OCCURRED*
30     16853720     20100211     ACTIVE
30     17965764     20100319     OCCURRED
30     16386708     20100211     COMPLETE
30     16804451     20100211     OCCURRED
30     15977897     20100121     ACTIVE
Edited by: sarvan on Aug 12, 2011 7:16 AM

try this [Not fully tested]
SQL> select * from v$version;
BANNER
Oracle Database 11g Enterprise Edition Release 11.2.0.2.0 - 64bit Production
PL/SQL Release 11.2.0.2.0 - Production
CORE    11.2.0.2.0      Production
TNS for Linux: Version 11.2.0.2.0 - Production
NLSRTL Version 11.2.0.2.0 - Production
Elapsed: 00:00:00.00
SQL> SELECT *
2    FROM (SELECT order_id, event_id, evnt_date, state_code, next_evntid,
3                 ROW_NUMBER () OVER (PARTITION BY event_id ORDER BY NULL)
4                                                                       AS rownm
5            FROM (WITH t AS
6                       (SELECT 10 AS order_id, 16937555 AS event_id,
7                               20100212 AS evnt_date, 'COMPLETE' AS state_code
8                          FROM DUAL
9                        UNION ALL
10                        SELECT 10, 16937555, 20100212, 'ACTIVE'
11                          FROM DUAL
12                        UNION ALL
13                        SELECT 10, 16308004, 20100129, 'OCCURRED'
14                          FROM DUAL
15                        UNION ALL
16                        SELECT 10, 16131904, 20100125, 'ACTIVE'
17                          FROM DUAL
18                        UNION ALL
19                        SELECT 10, 16270684, 20100128, 'OCCURRED'
20                          FROM DUAL
21                        UNION ALL
22                        SELECT 20, 17134164, 20100223, 'COMPLETE'
23                          FROM DUAL
24                        UNION ALL
25                        SELECT 20, 17134164, 20100223, 'ACTIVE'
26                          FROM DUAL
27                        UNION ALL
28                        SELECT 20, 17134164, 20100223, 'STARTED'
29                          FROM DUAL
30                        UNION ALL
31                        SELECT 20, 15479131, 20100105, 'OCCURRED'
32                          FROM DUAL) -- End of test data
33                  SELECT order_id, event_id, evnt_date, state_code,
34                         LEAD (event_id, 1, 0) OVER (PARTITION BY order_id ORDER BY NULL)
35                                                                 AS next_evntid
36                    FROM t)
37           WHERE event_id = next_evntid)
38   WHERE rownm <= 2
39 /
ORDER_ID   EVENT_ID EVNT_DATE STATE_CO NEXT_EVNTID      ROWNM
        10   16937555   20100212 COMPLETE    16937555          1
        20   17134164   20100223 COMPLETE    17134164          1
        20   17134164   20100223 ACTIVE      17134164          2
Elapsed: 00:00:00.00
SQL> PS - You should seriously think about ordering the data before you do this kind off operations.
Edited by: Sri on Aug 12, 2011 9:12 AM

Idoc to file scenario general query

Hi All,
I know about the basic settings needed for the Idoc to file scenario.
My query is related the idoc /ALE settings.
I have a scenatio wrking in my system where we have configured the port for B type partner and also the port as XML-HTTP
I can not see any DM(BD64) created for it basically.I want to know that in what scenarios the BD64 configuration is required.
Or it is the case that if we define the BD64 then we do not need to configure we20 and we21, it will be auto configured?
Please let me know the use cases for these different configurations if any.
Thanks in advance.
Regards,
Rahul Kulkarni

Hi Rahul,
BD64 uses mostly IDoc to IDoc scenarios because the IDoc consults the distribution model and determines whether any filter objects are specified for a receiver.
WE20 and WE21 (Partner Profile,Port Definition) we have to define in any scenarios where ECC included.
Thanks,
Seshu.

General query regarding the implementation of a Flash Streaming Server

Hey lads,
I'm looking at setting up an Adobe Flash Streaming Server at
work. It will be used primarily to stream training videos across
our Intranet - we have ~1000 users at the moment, with the capacity
to expand to ~2000). We are looking at both filming with a camera
and screencasting (using either
Camtasia or Adobe
Captivate) and then encoding to .FLV
Camtasia can encode directly to FLV, however I'm not too sure
of the quality, so the Adobe Flash Media Live Encoder looks like a
good option too. For the time being we will just be concentrating
on the screencasting though.
The good thing is i can download a development server for
free from Adobe which gives me complete access with just a
restriction on the amount of users that can access the content
(10).
From what i can gather the (simple) way it works (and that
I want it to work) is that you have a standard template SWF
file containing relevant actionscript (sitting in most probably in
a html/php/asp/ect file on the webserver), which then calls the FLV
file from the flash server and streams the content to the clients
PC. I assume i will pass a variable to the SWF file (the video ID i
want), and that will then be used to call the corresponding FLV
file.
So effectively a YouTube website, i.e:
watch?
v=WjK6wNIWzts
I just wanted to know if anyone had any general tips on how i
should approach this? Obviously quality vs file size will be a big
point - deciding on a standard FLV format will be a mission,
however i do have the benefit of an internal network so can anyone
recommend some good encoding tips? I was actually thinking of
encoding a low quality and high quality version, but i don't think
that's really sustainable/effective.
There is a bit of documentation around for this, but they
tend not to answer the basic questions. Am i heading in the right
direction?
Thanks for your help :)

If you are not able to find older version of connector on Metalink then raise an SR with Oracle and give proper Business Justification for the same.

Problem writing a sql query

I am a Java Developer and new to database.
Database Table:
Table 'Authors' with fields 'Author'(Varchar2), 'Code'(Varchar2), 'Id'(Varchar2 ~ Not a primary key, multiple records can have same value) and 'Date_Sequence'(Number). The possible values for 'Code' could be either 'Yes' or 'No'.
Approach:
I can select a list of authors, say AYES AUTHORS with code 'Yes' and a list of authors, say NOES AUTHORS with code 'No' from my web page.
When I submit the request, I need to retrieve all the Id's, where AYES Authors voted 'Yes' and NOES Authors voted 'No'.
I have come up with the following query:
SELECT DISTINCT ID FROM AUTHORS WHERE AUTHOR IN ('my selected AYES list') AND CODE = 'Yes'
INTERSECT
SELECT DISTINCT ID FROM AUTHORS WHERE AUTHOR IN ('my selected NOES list') AND CODE = 'No'
Problem:
Now the problem here being, I have to retrieve the id's where the 'Date Sequence' is same for 'Yes' and 'No'. Not sure how to link the Date_Sequence for both the queries across the intersect clause. Any clue is highly appreciated.
Thanks.

Not sure if i understand what the logic behind it is, but try this..
WITH AUTHORS AS (SELECT 'AYES1' AUTHOR, 'Yes' CODE, '1' ID, TRUNC(SYSDATE) DATE_SEQ FROM DUAL
                 UNION ALL
                 SELECT 'AYES2' AUTHOR, 'Yes' CODE, '2' ID, TRUNC(SYSDATE)-1 DATE_SEQ FROM DUAL
                 UNION ALL
                 SELECT 'AYES3' AUTHOR, 'No' CODE, '3' ID, TRUNC(SYSDATE) DATE_SEQ FROM DUAL
                 UNION ALL
                 SELECT 'NOES1' AUTHOR, 'No' CODE, '1' ID, TRUNC(SYSDATE) DATE_SEQ FROM DUAL
                 UNION ALL
                 SELECT 'NOES2' AUTHOR, 'No' CODE, '2' ID, TRUNC(SYSDATE) DATE_SEQ FROM DUAL
SELECT AYES.ID, AYES.DATE_SEQ
FROM AUTHORS AYES,
       AUTHORS NOES
WHERE AYES.AUTHOR IN ('AYES1', 'AYES2', 'AYES3') AND AYES.CODE='Yes'
   AND NOES.AUTHOR IN ('NOES1', 'NOES2', 'NOES3') AND NOES.CODE='No'
   AND AYES.ID=NOES.ID
   AND AYES.DATE_SEQ=NOES.DATE_SEQ
ID         DATE_SEQ
1          22.10.08
SQL>

Please help me in writing a simple query....

Hi,
I want a single query to retrieve the data as shown below.....
i have two tables
master(id primary key, name);
child( childid primary key, name, id)
here id referes the master(id)
let us say the data in the two tables like
master
id name
1 master1
2 master2
3 master3
child
childid name id
1 child1 1
2 child2 1
3 child3 2
4 child4 3
5 child5 3
6 child6 3
Now the result i want is like
id name parentid
1 master1 0
1 child1 1
2 child2 1
2 master2 0
3 child3 2
3 master3 0
4 child4 3
5 child5 3
6 child6 3
thanks in advance,
regards,
khaleel

SELECT   master.id,
         master.name,
         0 AS parentid,
         master.name AS mastername,
         NULL AS childname,
         'parent' AS type
FROM     master
UNION
SELECT   child.childid AS id,
         child.name,
         child.id AS parentid,
         master.name AS mastername,
         child.name AS childname,
         'child' AS type
FROM     child,
         master
WHERE    child.id = master.id
ORDER BY mastername,
         type DESC,
         childname
        ID NAME         PARENTID MASTERNAME CHILDNAME TYPE
         1 master1             0 master1               parent
         1 child1              1 master1    child1     child
         2 child2              1 master1    child2     child
         2 master2             0 master2               parent
         3 child3              2 master2    child3     child
         3 master3             0 master3               parent
         4 child4              3 master3    child4     child
         5 child5              3 master3    child5     child
         6 child6              3 master3    child6     child
9 rows selected.

WRITING A GENERAL qUERY

Similar Messages

Maybe you are looking for