Connect by prior in disco

Hi
I have a table in Oracle which is a recursive (pigs ear) hierarchy. I want to be able to use this in disco and "wander" down the hierarchies using connect by prior and start with. Can this be done?

Not sure if you're referring to a specific table.
However, I have create a report in the past that simply pointed to a database view we created that used the connect by idea and all worked well.
So, if you have a database view that returns what you're needing then no problem.

Similar Messages

SQL with connect by prior running for a long time

Hi,
We are using Oracle 10g. Below is a cursor sql which is having performance issues. The pAccountid is being passed from the output of a different cursor. But this cursor sql is running for a long time. Could you please help me in tuning this sql. I believe the subquery with connect by prior is causing the trouble.
The TRXNS is a huge table which is not partitioned. The query is forced to use the index on the accountid of the TRXNS table.
The accountlink table has 20,000 records and the TRXNStrack table has 10,000 records in total.
This sql executes for 200,000 pAccountids and runs for more than 8 hours.
SELECT /*+ INDEX(T TRXNS_ACCOUNTID_NIDX) */ AL.FROMACCOUNTID oldaccountid ,
                                A.ACCOUNTNUM oldaccountnum,
                               T.TRXNSID,
                               T.TRXNSTYPEID,
                               T.DESCRIPTION ,
                               T.postdt,
                               T.TRXNSAMT
                    FROM
                    ACCOUNTLINK AL,
                    TRXNS T,
                    ACCOUNT A
                   WHERE AL.TOACCOUNTID IN
                                                         (SELECT TOACCOUNTID FROM ACCOUNTLINK START WITH TOACCOUNTID = pAccountid
                                                                                                                     CONNECT BY PRIOR FROMACCOUNTID = TOACCOUNTID)
                        AND AL.FROMACCOUNTID = T.ACCOUNTID
                        AND A.ACCOUNTID = AL.FROMACCOUNTID
AND NOT EXISTS (select 1 from TRXNStrack trck where trck.TRXNSid = t.TRXNSid AND TRXNSTrackReasonid = 1)
                        AND T.postdt > A.CLOSEDATE
                        AND T.postdt >= sysdate-2
                        AND T.postdt <= sysdate;
Create script for trxn table:
CREATE TABLE SP.TRXNS
TRXNSID      NUMBER(15) CONSTRAINT "BIN$rpIQEeyLDfbgRAAUT4DEnQ==$0" NOT NULL,
ACCOUNTID    NUMBER(15) CONSTRAINT "BIN$rpIQEeyMDfbgRAAUT4DEnQ==$0" NOT NULL,
STATEMENTID NUMBER(15),
TRXNSTYPEID NUMBER(15),
DESCRIPTION VARCHAR2(80 BYTE),
postdt     DATE,
TRXNSAMT     NUMBER(12,2),
TRXNSREQID   NUMBER(15),
LASTUPDATE   DATE,
SOURCEID     NUMBER(15),
HIDE         VARCHAR2(1 BYTE)
TABLESPACE SO_TRXN_DATA
RESULT_CACHE (MODE DEFAULT)
PCTUSED    40
PCTFREE    10
INITRXNS   2
MAXTRXNS   255
STORAGE    (
            INITIAL          50M
            NEXT             1M
            MINEXTENTS       1
            MAXEXTENTS       UNLIMITED
            PCTINCREASE      0
            FREELISTS        8
            FREELIST GROUPS 1
            BUFFER_POOL      DEFAULT
            FLASH_CACHE      DEFAULT
            CELL_FLASH_CACHE DEFAULT
LOGGING
NOCOMPRESS
NOCACHE
NOPARALLEL
MONITORING;
CREATE INDEX SP.TRXNS_ACCOUNTID_NIDX ON SP.TRXNS
(ACCOUNTID, postdt)
LOGGING
TABLESPACE SO_TRXN_INDEX
PCTFREE    10
INITRXNS   2
MAXTRXNS   255
STORAGE    (
            INITIAL          64K
            NEXT             1M
            MINEXTENTS       1
            MAXEXTENTS       UNLIMITED
            PCTINCREASE      0
            FREELISTS        1
            FREELIST GROUPS 1
            BUFFER_POOL      DEFAULT
            FLASH_CACHE      DEFAULT
            CELL_FLASH_CACHE DEFAULT
NOPARALLEL;
below is the executing plan for this sql taken from toad :
PLAN_ID
TIMESTAMP
OPERATION
OPTIONS
OBJECT_OWNER
OBJECT_NAME
OBJECT_ALIAS
OBJECT_INSTANCE
OBJECT_TYPE
OPTIMIZER
SEARCH_COLUMNS
ID
PARENT_ID
DEPTH
POSITION
COST
CARDINALITY
BYTES
CPU_COST
IO_COST
TEMP_SPACE
ACCESS_PREDICATES
FILTER_PREDICATES
PROJECTION
TIME
QBLOCK_NAME
1121
9/10/2013 3:30
FILTER
1
0
1
1
NOT EXISTS (SELECT 0 FROM "TRXNSTRACK" "TRCK" WHERE "TRXNSTRACKREASONID"=1 AND "TRCK"."TRXNSID"=:B1)
AL."FROMACCOUNTID"[NUMBER,22], "T"."TRXNSID"[NUMBER,22], "T"."TRXNSTYPEID"[NUMBER,22], "T"."DESCRIPTION"[VARCHAR2,80], "T"."POSTDT"[DATE,7], "T"."TRXNSAMT"[NUMBER,22], "A"."ACCOUNTNUM"[VARCHAR2,19]
SEL$5DA710D3
1121
9/10/2013 3:30
FILTER
2
1
2
1
SYSDATE@!-2<=SYSDATE@!
AL."FROMACCOUNTID"[NUMBER,22], "T"."TRXNSID"[NUMBER,22], "T"."TRXNSTYPEID"[NUMBER,22], "T"."DESCRIPTION"[VARCHAR2,80], "T"."POSTDT"[DATE,7], "T"."TRXNSAMT"[NUMBER,22], "A"."ACCOUNTNUM"[VARCHAR2,19]
1121
9/10/2013 3:30
NESTED LOOPS
3
2
3
1
(#keys=0) "AL"."FROMACCOUNTID"[NUMBER,22], "T"."TRXNSID"[NUMBER,22], "T"."TRXNSTYPEID"[NUMBER,22], "T"."DESCRIPTION"[VARCHAR2,80], "T"."POSTDT"[DATE,7], "T"."TRXNSAMT"[NUMBER,22], "A"."ACCOUNTNUM"[VARCHAR2,19]
1121
9/10/2013 3:30
NESTED LOOPS
4
3
4
1
5
1
119
3989858
4
(#keys=0) "AL"."FROMACCOUNTID"[NUMBER,22], "T"."TRXNSID"[NUMBER,22], "T"."TRXNSTYPEID"[NUMBER,22], "T"."DESCRIPTION"[VARCHAR2,80], "T"."POSTDT"[DATE,7], "T"."TRXNSAMT"[NUMBER,22], "A".ROWID[ROWID,10]
1
1121
9/10/2013 3:30
NESTED LOOPS
5
4
5
1
4
1
90
3989690
3
(#keys=0) "AL"."FROMACCOUNTID"[NUMBER,22], "T"."TRXNSID"[NUMBER,22], "T"."TRXNSTYPEID"[NUMBER,22], "T"."DESCRIPTION"[VARCHAR2,80], "T"."POSTDT"[DATE,7], "T"."TRXNSAMT"[NUMBER,22]
1
1121
9/10/2013 3:30
HASH JOIN
SEMI
6
5
6
1
3
2
54
3989094
2
AL."TOACCOUNTID"="TOACCOUNTID"
(#keys=1) "AL"."FROMACCOUNTID"[NUMBER,22]
1
1121
9/10/2013 3:30
INDEX
FULL SCAN
SP
ACCOUNTLINK_AK1
AL@SEL$1
INDEX (UNIQUE)
ANALYZED
7
6
7
1
1
18
252
107
1
AL."FROMACCOUNTID"[NUMBER,22], "AL"."TOACCOUNTID"[NUMBER,22]
1
SEL$5DA710D3
1121
9/10/2013 3:30
VIEW
SYS
VW_NSO_1
VW_NSO_1@SEL$5DA710D3
11
VIEW
8
6
7
2
2
18
234
107
1
TOACCOUNTID[NUMBER,22]
1
SEL$683B0107
1121
9/10/2013 3:30
CONNECT BY
NO FILTERING WITH START-WITH
9
8
8
1
TOACCOUNTID=PRIOR "FROMACCOUNTID"
TOACCOUNTID=56354162
TOACCOUNTID[NUMBER,22], "FROMACCOUNTID"[NUMBER,22], PRIOR NULL[22], LEVEL[4]
SEL$683B0107
1121
9/10/2013 3:30
INDEX
FULL SCAN
SP
ACCOUNTLINK_AK1
ACCOUNTLINK@SEL$3
INDEX (UNIQUE)
ANALYZED
10
9
9
1
1
18
252
107
1
ACCOUNTLINK.ROWID[ROWID,10], "FROMACCOUNTID"[NUMBER,22], "TOACCOUNTID"[NUMBER,22]
1
SEL$3
1121
9/10/2013 3:30
TABLE ACCESS
BY INDEX ROWID
SP
TRXNS
T@SEL$1
2
TABLE
ANALYZED
11
5
6
2
1
1
63
298
1
T."TRXNSID"[NUMBER,22], "T"."TRXNSTYPEID"[NUMBER,22], "T"."DESCRIPTION"[VARCHAR2,80], "T"."POSTDT"[DATE,7], "T"."TRXNSAMT"[NUMBER,22]
1
SEL$5DA710D3
1121
9/10/2013 3:30
INDEX
RANGE SCAN
SP
TRXNS_ACCOUNTID_NIDX
T@SEL$1
INDEX
ANALYZED
2
12
11
7
1
1
1
224
1
AL."FROMACCOUNTID"="T"."ACCOUNTID" AND "T"."POSTDT">=SYSDATE@!-2 AND "T"."POSTDT"<=SYSDATE@!
T.ROWID[ROWID,10], "T"."POSTDT"[DATE,7]
1
SEL$5DA710D3
1121
9/10/2013 3:30
INDEX
UNIQUE SCAN
SP
ACCOUNT_PK
A@SEL$1
INDEX (UNIQUE)
ANALYZED
1
13
4
5
2
1
1
90
1
A."ACCOUNTID"="AL"."FROMACCOUNTID"
A.ROWID[ROWID,10]
1
SEL$5DA710D3
1121
9/10/2013 3:30
TABLE ACCESS
BY INDEX ROWID
SP
ACCOUNT
A@SEL$1
3
TABLE
ANALYZED
14
3
4
2
1
1
29
168
1
A."CLOSEDATE"<SYSDATE@! AND "T"."POSTDT">"A"."CLOSEDATE"
A."ACCOUNTNUM"[VARCHAR2,19]
1
SEL$5DA710D3
1121
9/10/2013 3:30
INDEX
RANGE SCAN
SP
TRXNSTRACK_TRXNSID_NIDX
TRCK@SEL$6
INDEX
ANALYZED
2
15
1
2
2
1
1
10
73
1
TRCK."TRXNSID"=:B1 AND "TRXNSTRACKREASONID"=1
TRCK."TRXNSID"[NUMBER,22], "TRXNSTRACKREASONID"[NUMBER,22]
1
SEL$6
Please help me in debugging this thanks!

Hi,
Thanks for your thought on this subject. Below is the trace info that I got from the DBA
SQL ID: d0x879qx2zgtz Plan Hash: 4036333519
SELECT /*+ INDEX(T TRXNS_ACCOUNTID_NIDX) */ AL.FROMACCOUNTID OLDACCOUNTID ,
A.ACCOUNTNUM OLDACCOUNTNUM, T.TRXNSID, T.TRXNSTYPEID, T.DESCRIPTION ,
T.POSTDT, T.TRXNSAMT
FROM
ACCOUNTLINK AL, TRXNS T, ACCOUNT A WHERE AL.TOACCOUNTID IN (SELECT
TOACCOUNTID FROM ACCOUNTLINK START WITH TOACCOUNTID = :B3 CONNECT BY PRIOR
FROMACCOUNTID = TOACCOUNTID) AND AL.FROMACCOUNTID = T.ACCOUNTID AND
A.ACCOUNTID = AL.FROMACCOUNTID AND NOT EXISTS (SELECT 1 FROM TRXNSTRACK
TRCK WHERE TRCK.TRXNSID = T.TRXNSID AND TRXNSTRACKREASONID = :B4 ) AND
T.POSTDT > A.CLOSEDATE AND T.POSTDT >= :B2 AND T.POSTDT <= :B1
call     count       cpu    elapsed       disk      query    current        rows
Parse        0      0.00       0.00          0          0          0           0
Execute 17160      2.10       1.87          0          0          0           0
Fetch    17160   7354.61    7390.86     169408    5569856 883366791           0
total    34320   7356.71    7392.74     169408    5569856 883366791           0
Misses in library cache during parse: 0
Optimizer mode: CHOOSE
Parsing user id: 38     (recursive depth: 1)
SQL ID: gs89hpavb4cts Plan Hash: 3415795327
SELECT A.ACCOUNTID, C.MEMBERID, A.PROGRAMID, A.ACCOUNTNUM
FROM
CUSTOMER C, CUSTOMERACCOUNT CA, ACCOUNT A, PROGRAMPARAMVALUE PPV,
BATCHPROCESSPROGRAM BP WHERE A.PROGRAMID = BP.PROGRAMID AND A.PROGRAMID =
PPV.PROGRAMID AND A.ACCOUNTID = CA.ACCOUNTID AND CA.PERSONID = C.PERSONID
AND PPV.PARAMID = :B2 AND PPV.VALUE = 'Y' AND BP.PROCESSID = :B1 AND BP.RUN
= 'Y' AND A.ACCOUNTTYPEID = 4 AND A.ACCOUNTSTATUSID = 1 AND C.MEMBERID IS
NOT NULL
call     count       cpu    elapsed       disk      query    current        rows
Parse        0      0.00       0.00          0          0          0           0
Execute      0      0.00       0.00          0          0          0           0
Fetch      172     13.14     115.34      80826     278650          0       17200
total      172     13.14     115.34      80826     278650          0       17200
Misses in library cache during parse: 0
Parsing user id: 38     (recursive depth: 1)
OVERALL TOTALS FOR ALL NON-RECURSIVE STATEMENTS
call     count       cpu    elapsed       disk      query    current        rows
Parse        0      0.00       0.00          0          0          0           0
Execute      0      0.00       0.00          0          0          0           0
Fetch        0      0.00       0.00          0          0          0           0
total        0      0.00       0.00          0          0          0           0
Misses in library cache during parse: 0
OVERALL TOTALS FOR ALL RECURSIVE STATEMENTS
call     count       cpu    elapsed       disk      query    current        rows
Parse        0      0.00       0.00          0          0          0           0
Execute 17160      2.10       1.87          0          0          0           0
Fetch    17332   7367.75    7506.21     250234    5848506 883366791       17200
total    34492   7369.85    7508.09     250234    5848506 883366791       17200
Misses in library cache during parse: 0
    2 user SQL statements in session.
    0 internal SQL statements in session.
    2 SQL statements in session.
Trace file: svoprod_ora_12346.trc
Trace file compatibility: 11.1.0.7
Sort options: default
       1 session in tracefile.
       2 user SQL statements in trace file.
       0 internal SQL statements in trace file.
       2 SQL statements in trace file.
       2 unique SQL statements in trace file.
   66499 lines in trace file.
    7516 elapsed seconds in trace file.

Connect by Prior in Oracle 8i question

Need help with building tree-structured query in Oracle 8i (Forms 6i front end if it matters).
Sample structure and data:
CREATE TABLE table_list
my_code NUMBER,
my_level NUMBER,
my_description VARCHAR2(60)
CREATE TABLE table_content
my_code NUMBER,
term_code NUMBER,
term_category VARCHAR2(5)
INSERT into table_list values (101, 1, 'building');
INSERT into table_list values (102, 2, 'flat');
INSERT into table_list values (103, 3, 'living room');
INSERT into table_list values (104, 3, 'bedroom');
INSERT into table_list values (105, 3, 'bathroom');
commit;
INSERT into table_content values (101, 102, 'Sub');
INSERT into table_content values (102, 103, 'Sub');
INSERT into table_content values (102, 104, 'Sub');
INSERT into table_content values (102, 105, 'Sub');
commit;
Need to display data in the following order:
101 'building' --level one
     102 'flat' --level two
          105 'bathroom' --level three
          104 'bedroom' --level three
          103 'living room' --level three
*(note alphabetical order in level three)*
Looks like Oracle 8i does not support table joins for CONNECT BY PRIOR. Please advise!

And you are correct, it does not display level 1. Do
you think this is this a database structure problem
or it could be corrected via query modification?No and yes. It's not a "structure" problem, its a data problem (as I explained above). You always fix data problems by making queries overly complex (hint, hint, I don't suggest doing it this way).
select tl.my_level, tl.my_description , LPAD(' ',3*(my_level-1)) || substr(tl.my_description, 1, 30) description
from table_list tl,
(select term_Code from table_content
start with my_code in (select my_code from table_list where my_level=1)
connect by prior term_code = my_code
) x
where x.term_code = tl.my_code
union all
select my_level, my_description
from table_list
where my_level = 1
order by 1, 2
better to add a row to table_content with my_code=null, term_code=101.
otherwise you'll be pulling stupid nonsense like above with every query from here on out.

How to use simple SQL instead of Connect By Prior

Currently, I am using "connect by prior" query in the application, but for reason I don�t want to use this connect by query so can any one please tell how does I get the same result by using SQL, I tried this by using procedure but unable to get the same result, specially LEVEL of the tree.
So please tell, how would I get the correct data.
Thanks in advance,
AMIT.

Hi,
Whenever you have a question, it helps to post:
(1) The version of Oracle (and any other relevant software) you're using
(2) A little sample data (just enough to show what the problem is) from all the relevant tables
(3) The results you want from that data
(4) Your best attempt so far (formatted) I don't believe the unformated code you posted is what you're really running, since it has a syntax error ("... WHERE START WITH ..."). Please post code that really works with the sample data you posrted.
(5) The full error message (if any), including line number
Executable SQL statements (like "CREATE TABLE AS ..." or "INSERT ..." statements) are best for (2).
If you can present your problem using commonly available tables (for example, scott.emp, which contains a hierarchy), then you can omit (2).
Formatted tabular output is okay for (3). Type these 6 characters
{code}
(small letters only, inside curly brackets) before and after the tabular text, to preserve spacing.
As Alex said, why don't you want to use CONNECT BY?
Are you getting the correct results now, but just looking for a different way of getting them?
Depending on your exact requirements, you could write a PLSQL function that mimics LEVEL. Don't expect it to be fast.
Nested Sets is a completely different way of modeling trees.
Some things are much easier with Nested Sets than they are using the Adjacency Model (the parent-child model that uses CONNECT BY).
But some things are much harder with Nested Sets, and LEVEL is one of them.

How to get the root node for a child using connect by prior

Hi,
I searched at many places to do this but not able to get the exact o/p sp posting my question here.
I have a table with parent_id and child_id columns. The levels of this parent_child can be many but my aim is to find the ultimate parent ( dont know the right term)
like if I have a child_node= xyz and its parent= pqr and its parents=lmn which might be ultimate parent which doesnt have any further parent.
So if i start with child_code= "xyz" then i should get the parent as "lmn" and not the immediate parent "pqr".
Please help.
Thanks,
Aashish

To find the Parent on emp table:
select empno,ename,level
from emp
where level=3
start with empno=7934
connect by empno= prior mgrTo find the Child on emp table:
select empno,ename,level
from emp
where level=1
start with empno=7934
connect by prior empno= mgrBut you need to know the level or position to specify the level of parent or child. But there are some other possible options available if we can know the requirement properly.
Edited by: Vasista on Jan 11, 2011 2:27 AM
Edited by: Vasista on Jan 11, 2011 2:30 AM

Query tuning for query using connect by prior

I have written following query to fetch the data. The query is written in this format because there are multiple rows, which make one recrd and we need to bring that record into one row.
For one CAT(commented here), this query takes around 4 minutes and fetches 6900 records but when it runs for 3 CAT, it takes 17 mins.
I want to tune this as this has to run for 350 CAT values.
It is doing FTS on the main table. I tried to use different hints like PARALLEL, APPEND (in insert) but nothing worked.
The cost of the query is 51.
Any help/suggestions will be appreciated.
SELECT DISTINCT MIN(SEQ) SEQ,
PT, APP, IT, LR, QT,CD, M_A_FLAG,
STRAGG(REM) REM, -- aggregates the data from different columns to one which is parent
CAT
FROM (WITH R AS (SELECT CAT, SEQ, PT, M_A_FLAG, IT, LR,QT,CD, REM, APP
FROM table1
WHERE REC = '36' AND M_A_FLAG = '1'
--AND CAT = '11113')
SELECT CAT, SEQ,
CONNECT_BY_ROOT PT AS PT,
CONNECT_BY_ROOT APP AS APPL,
M_A_FLAG,
CONNECT_BY_ROOT IT AS IT,
CONNECT_BY_ROOT LR AS LR,
CONNECT_BY_ROOT QT AS QT,
CONNECT_BY_ROOT CD AS CD,
REM
FROM R A
START WITH PT IS NOT NULL
CONNECT BY PRIOR SEQ + 1 = SEQ
AND PRIOR CAT = CAT
AND PT IS NULL)
GROUP BY PT, APP, IT,LR, QT, CD, M_A_FLAG, CAT
ORDER BY SEQ;
Thanks.
Edited by: user2544469 on Feb 11, 2011 1:12 AM

The following threads detail the approach and information required.
Please gather relevant info and post back.
How to post a SQL tuning request - HOW TO: Post a SQL statement tuning request - template posting
When your query takes too long - When your query takes too long ...

Connect by prior in Oracle 11g

I am upgrading an application from 9i to 11g and I've been told that the connect by prior sql is broken. While I am waiting for the example and the error message, is 11g pickier about looping than 9i was? Can someone point me to some documentation on 10 or 11 changes to hierarchical queries?
Thanks in advance.
Jim

Hi, Jim,
user509659 wrote:
I am upgrading an application from 9i to 11g and I've been told that the connect by prior sql is broken. What do you mean? Post a link, or at least a more complete description of the problem.
While I am waiting for the example and the error message, is 11g pickier about looping than 9i was? Can someone point me to some documentation on 10 or 11 changes to hierarchical queries?There were a lot of improvements to CONNECT BY in Oracle 10, including CONNECT BY NOCYCLE (for working with loops), the CONNECT_BY_ROOT operator, and pseudo-columns CONNECT_BY_ISCYCLE and CONNECT_BY_ISLEAF. Look them up in the SQL Language manual for your version.
I don't know of any improvements or changes to CONNECT BY in Oracle 11.
Oracle 11.2 has recursive WITH clauses, that can do everything that CONNECT BY does, and more. Some things in 11.2 are simpler and more efficient using CONNECT BY, others are better done using recursive WITH clauses.
Everything that CONNECT BY did in earlier versions, it does in Oracle 11, as well. If you have a CONNECT BY query that works in Oracle 9, it should work in Oracle 11, without any changes, though there might be better ways to get the same results in Oracle 11.
Whenever you have a question, post youir code, and whatever is necessary to run it, including CREATE TABLE and INSERT statments for a little sample data. Post the results you want from that sample data, and an explanation of how you get those results from that data.
Always say which version of Oracle you're using, e.g. 11.2.0.2.0. The difference between 11.1 and 11.2 may be very significant in this case.
See the forum FAQ {message:id=9360002}

Help me in understand connect By Prior

Hi ,
please help me in understand connect By Prior
I did a sample example , but unale to follow , please explain
How did it understand that KING shuld be displaed first .
On wht basis the results are shown here ?
SELECT empno,
ename,
job,
mgr,
hiredate,
level
FROM   emp
START WITH mgr IS NULL
CONNECT BY PRIOR empno = mgr
7839     KING PRESIDENT      17-Nov-81     1
7566     JONES MANAGER 7839     2-Apr-81     2
7788     SCOTT ANALYST 7566     19-Apr-87     3
7876     ADAMS CLERK 7788     23-May-87     4
7902     FORD ANALYST 7566     3-Dec-81     3
7369     SMITH CLERK 7902     17-Dec-80     4
7698     BLAKE MANAGER 7839     1-May-81     2
7499     ALLEN SALESMAN 7698     20-Feb-81     3
7521     WARD SALESMAN 7698     22-Feb-81     3
7654     MARTIN SALESMAN 7698     28-Sep-81     3
7844     TURNER SALESMAN 7698     8-Sep-81     3
7900     JAMES CLERK 7698     3-Dec-81     3
7782     CLARK MANAGER 7839     9-Jun-81     2
7934     MILLER CLERK 7782     23-Jan-82     3

Hi,
user10503747 wrote:
Hi ,
please help me in understand connect By Prior
I did a sample example , but unale to follow , please explain
How did it understand that KING shuld be displaed first .
On wht basis the results are shown here ?In a CONNECT BY query (without an ORDER BY clause), if x is an ancestor of y (that is, if x is the parent of y, or x is the parent of the parent of y, and so on), then y will be displayed after x, but before any other row that does not have x as its ancestor. (That is also the order in which ROWNUM will be assigned. This applies only to the query in which CONNECT BY is done. If you use the results set as a sub-query, the super query may cause the results to be re-arranged. An ORDER BY clause always takes precedence.)
In a CONNECT BY query, every row in the result set either:
(a) satisfies the START WITH condition, or
(b) is a descendant of some row in (a); that is, some row that satifies the START WITH condition is its ancestor.
Since ancestors always come before their descendants, the first row in the result set must be a row that satisfied the START WITH condition. In your example, the row with ename='KING' was the only row that satisfied the condition "START WITH mgr IS NULL", and all the other rows are its descendants, so the row with ename='KING' must come first, as APC said.

Oracle Virtual Machine Connect by Prior Problem

Hi Friends ,
In my company we have virtualization. this is new in our organization.
oracle image have been taken and put in to virtual machine.(vmware).
after that, procedure which has "connect by prior" start to create wrong output.
please help us on what to do to correct problem given above.
(my oracle version 10g)

JAVAMAN2 wrote:
simple sample:
SELECT ROWNUM+2000-1 AS YIL FROM DUAL CONNECT BY ROWNUM <= 2010-2005 ORDER BY ROWNUM+2000-1 DESC;output on virtual machine
2009,2008,2007,2006,2005
output on real world
2010,2009,2008,2007,2006,2005If that is really the output from that query, then both your virtual machine and your "real world" are broken. On all 4 versions of Oracle I have available, I get:
SQL> SELECT ROWNUM + 2000 - 1 AS YIL
2 FROM DUAL
3 CONNECT BY ROWNUM <= 2010 - 2005
4 ORDER BY ROWNUM + 2000 - 1 DESC;
       YIL
      2004
      2003
      2002
      2001
      2000Which is what I would expect given that 2010 - 2005 = 5.
Perhaps if you showed an actual cut and paste from a sqlplus session on each machine someone might be able to help, however, I'm inclined to agree with sb92075 that it is a data issue in your real query.
John

Connect By Prior Help

Hi all,
In Connect By Prior, How can i avoid, if some parent dont have child means, it should not come in the results.......
Regards,
Fame

A "parent" without a child will not show up as a parent, but it may show up as a child.
When that happens, CONNECT_BY_ISLEAF will be 1.
So, filter on CONNECT_BY_ISLEAF = 0.select empno, ename, job, mgr from scott.emp
where connect_by_isleaf = 0
start with mgr is null
connect by mgr = prior empno;
     EMPNO ENAME      JOB              MGR
      7839 KING       PRESIDENT
      7566 JONES      MANAGER         7839
      7788 SCOTT      ANALYST         7566
      7902 FORD       ANALYST         7566
      7698 BLAKE      MANAGER         7839
      7782 CLARK      MANAGER         7839Edited by: Stew Ashton on Jan 17, 2013 1:34 PM

Connect by prior problem with order by

Hi,
I and using connect by prior within a query on Oracle 8 and would like to order the results within the parent levels.
All the documentation that I have read shows that in Oracle 9i there is an option to say order siblings by which looks like what I need, but this does not work on Oracle 8.
Can anyone tell me how I can order the children within the parents without changing the tree structure?
I have also tried SYS_CONNECT_BY_PATH and I just get an error saying that it is an invalid column name,

Karen, see here for a dicussion on how to order the siblings in a pre-9i environment:
http://asktom.oracle.com/pls/ask/f?p=4950:8:::::F4950_P8_DISPLAYID:9212348049

CONNECT BY PRIOR on 9i

This 9i SQL Reference indicates that it is possible to combine CONNECT BY PRIOR with a join predicate.
In 8i I get the following error:
ORA-01437 cannot have join with CONNECT BY
Cause: A join operation was specified with a CONNECT BY clause. If a CONNECT BY clause is used in a SELECT statement for a tree-structured query, only one table may be referenced in the query.
Action: Remove either the CONNECT BY clause or the join operation from the SQL statement.
I can't find documentation to test the obvious hypothesis - comments?

My alternative solutions.
***Sample***
ID PrevID
1    null
2       1
3       1
4       3
***output which we want***
ID PrevID isLeaf
1    null       0
2       1       1
3       1       0
4       3       1
***DDL***
create table CloneIsLeaf as
select 1 as ID,null as PrevID from dual
union select 2,1 from dual
union select 3,1 from dual
union select 4,3 from dual;
--method1
select ID,PrevID,
case when exists(select 1 from CloneIsLeaf b
                  where a.ID = b.PrevID) then 0 else 1 end as isLeaf
from CloneIsLeaf a
Start With ID = 1
connect by prior ID = PrevID;
--method2
select ID,PrevID,
case when Level < Lead(Level) over(order by RowNum)
     then 0 else 1 end as isLeaf
from CloneIsLeaf
Start With ID = 1
connect by prior ID = PrevID;
--method3
select ID,PrevID,
case when LV < Lead(LV) over(order by RowNum)
     then 0 else 1 end as isLeaf
from (select ID,PrevID,Level as LV
        from CloneIsLeaf
      Start With ID = 1
      connect by prior ID = PrevID
      order siblings by ID);
--method4
select ID,PrevID,
case when LV < Lead(LV) over(order by Row_Num)
     then 0 else 1 end as isLeaf
from (select ID,PrevID,LV,RowNum as Row_Num
        from (select ID,PrevID,Level as LV
                from CloneIsLeaf
              Start With ID = 1
              connect by prior ID = PrevID
              order siblings by ID));on method2 and method3 and method3,
I used that "Hierarchical Queries" is depth-first search (http://en.wikipedia.org/wiki/Depth-first_search)
http://download-west.oracle.com/docs/cd/B19306_01/server.102/b14200/queries003.htm#i2053935
my site :-)
http://www.geocities.jp/oraclesqlpuzzle/10-149.html

Connect by prior performance

In our application, we have a lot of 'trees' in our data (parent/child relationships within a table). We rely heavily on 'connect by prior' to search through the trees, to find for example values that are defined in one of the parents ('inheritance').
Lately however, we've been eliminating connect by priors wherever we can lately, because performance is dramatic. No matter what indexes we use, we keep getting full tablescans on large tables because of the connect by priors. Views with 2 connect by priors are deadly for performance.
What is a good way to optimize connect by priors?
As an example of why we use connect by prior, consider this:
table A
ID   PARENT_ID   NAME    LENGTH
1    NULL        root    10
2    1           child   NULLWe consider record 2 'derived' from record 1, inheriting its length. The alternative for connect by prior is to populate the child length whenever the master length changes, and this is what we do in most situations now, but it's hell to keep track of this. If a record changes, you would still have to go down the tree to change all child lengths etc, so we rather use connect by prior.
Any ideas?

Ivo, would you please post the table definition (if different than your example), the actual list of indexes, a sample query, and the explain plan for that query?
I use hierarchical queries extensively on tables with millions of rows. They work very well if all the pieces are in place. Let's see if we can get your queries humming...

Connect by prior by speific order problem.

Oracle: 10.2.0.4
I have a table containing events backup.
i want to list the hierarchy within that backup oreder by the time_stamp for all levels.
create table BCK_EVENTS
bck_backup_id           NUMBER(9) default 0 not null,
event_id                   NUMBER(15) default 0 not null,
event_name              NVARCHAR2(100) default ' ' not null
time_stamp               NUMBER(9) default 0 not null,
parent_event            NUMBER(15) default 0 not null,
add constraint BCK_EVENTS_PK primary key (BCK_EVENT_ID, EVENT_ID); // event_id is not unique can be under one or more backup id'sthis is not a tree with one root, there is more than one event in the root level (level 1).
example:
Event Name      Time         level   <- time is numeric but for easier reading.
*Event A          10:00         1
   *Event C        10:30         2
     *Event B      11:17         3
*Event H          12:10         1
     *Event J       12:10         2
     *Event M      12:21         2
*Event Z          15:33         1
   *Event R        16:56          2
    *Event M      16:57          3
   *Event G        20:20         2What i tried was :
select lpad( '*', level*2 ) || event_id,event_name,time_stamp,parent_event,level from bck_events
     where bck_event_id=100031
     start with parent_event is null
    connect by prior event_id = parent_eventand there are two problems with it.
1. it's not ordered even when i added an Index (parent_event,time_stamp) and try to hint it.
2. it returns loads of multiple rows.
hope it's clear enough. I thank for any help.
Edited by: 973065 on Nov 25, 2012 8:04 AM
Edited by: 973065 on Nov 25, 2012 9:23 AM
Edited by: 973065 on Nov 25, 2012 9:29 AM
Edited by: 973065 on Nov 25, 2012 9:31 AM

Hi,
973065 wrote:
Oracle: 10.2
I have a table containing events backup.
i want to list the hierarchy within that backup oreder by the time_stamp for all levels.
create table BCK_EVENTS
bck_backup_id           NUMBER(9) default 0 not null,
event_id                   NUMBER(15) default 0 not null,
event_name              NVARCHAR2(100) default ' ' not null
time_stamp               NUMBER(9) default 0 not null,
parent_event            NUMBER(15) default 0 not null,
Thanks for posting the version number and the CREATE TABLE statement. Don't forget to post INSERT statements for your sample data.
add constraint BCK_EVENTS_PK primary key (BCK_EVENT_ID, EVENT_ID); // event_id is not unique can be under one or more backup id's
this is not a tree with one root, there is more than one event in the root level (level 1).Is it a forest, that is, a set of trees?
example:
Event Name      Time         level   <- time is numeric but for easier reading.
*Event A          10:00         1
*Event C        10:30         2
*Event B      11:17         3
*Event H          12:10         1
*Event J       12:10         2
*Event M      12:21         2
*Event Z          15:33         1
*Event R        16:56          2
*Event M      16:57          3
*Event G        20:20         2
That seems to be a forest, that is, every row has 0 or 1 parent, and no row is its own ancestor.
What i tried was :
select lpad( '*', level*2 ) || event_id,event_name,time_stamp,parent_event,level from bck_events
where bck_event_id=100031
start with parent_event is null
connect by prior event_id = parent_event
and there are two problems with it.
1. it's not ordered even when i added an Index (parent_event,time_stamp) and try to hint it.Depending on your data, you may just need to add
ORDER SIBLINGS BY time_stampat the end, after the CONNECT BY clause.
If you need rows sorted by time_stamp under their roots, but otherwise without regard to the hierarchy, then use CONNECT_BY_ROOT.
2. it returns loads of multiple rows.Again, it depends on your data. I'll bet you need something more in the CONNECT BY clause, but I can't tell what without some sample data and an exxplanation of how you gett the results you posted from that data. The fact that bck_event_id is part of the primary key makes me suspect that maybe bck_event_id needs to be somewhere in the CONNECT BY clause, but that's just a wild guess.
hope it's clear enough. I thank for any help.As mentioned before, see the forum FAQ {message:id=9360002}

CONNECT BY PRIOR statement issue

Hi,
I'm using the below select query to build a relationship b/w ID & parent_id, but HANA studio is throwing a syntax error. I don't see a problem around the syntax. Please throw some light's here.
SELECT * FROM "SIVA_TEST"."Core_Competency" START with "Competency_ID"=1
CONNECT BY PRIOR "Competency_ID" = "Parent_Competency_ID" ;
Error:
Could not execute 'SELECT * FROM "SIVA_TEST"."Core_Competency" START WITH "Competency_ID"='1' CONNECT BY PRIOR ...' in 170 ms 563 µs .
SAP DBTech JDBC: [257] (at 46): sql syntax error: incorrect syntax near "START": line 2 col 1 (at pos 46)
Table struct:
create column table "SIVA_TEST"."Core_Competency"("Competency_ID" integer,
"Competency_Name" varchar(200),
"Competency_description" varchar(1000),
"Parent_Competency_ID" integer,
"start_date" date,
primary key("Competency_ID"));
Thanks,
Siva.

Insert statement for the corresponding table..
INSERT INTO "SIVA_TEST"."Core_Competency" VALUES (1,'value 1',null,null,'2013-12-02');
INSERT INTO "SIVA_TEST"."Core_Competency" VALUES (2,'value 2',null,1,'2013-07-31');
INSERT INTO "SIVA_TEST"."Core_Competency" VALUES (3,'value 3',null,1,'2013-02-01');
INSERT INTO "SIVA_TEST"."Core_Competency" VALUES (4,'value 4',null,2,'2013-06-24');
INSERT INTO "SIVA_TEST"."Core_Competency" VALUES (5,'value 6',null,2,'2013-04-09');
INSERT INTO "SIVA_TEST"."Core_Competency" VALUES (6,'value 7',null,3,'2013-09-12');
INSERT INTO "SIVA_TEST"."Core_Competency" VALUES (7,'value 8',null,3,'2013-08-09');
INSERT INTO "SIVA_TEST"."Core_Competency" VALUES (8,'value 9',null,7,'2013-04-09');
INSERT INTO "SIVA_TEST"."Core_Competency" VALUES (9,'value 10',null,7,'2013-12-09');
INSERT INTO "SIVA_TEST"."Core_Competency" VALUES (10,'value 11',null,4,'2013-09-08');
INSERT INTO "SIVA_TEST"."Core_Competency" VALUES (11,'value 12',null,4,'2013-10-10');

Connect by prior in disco

Similar Messages

Maybe you are looking for