hierarchical SQL

Hello

I'm looking for help with a hierarchical query.

My table has five columns:

rank (number)
score (number)
name (varchar)
record_id (number)
ParentId (number)

A parentid is a record_id at a higher level in the hierarchy.
Rank and Score are used to control (tri) each level of the hierarchy.

The data looks like this:

grade - score - name - record_id - parentid
----------------------------------------------------------
1 - 1 - name1 - 101-0
1 - 2 - name2 - 102-0
1 - 3 - name3 - 103-0
2 - 1 - name4 - 104-101
2 - 2 - name5 - 105-101
2 - 1 - name6 - 106-102
2 - 2 - Marque7 - 107-102
2 - 1 - name8 - 108-103
2 - 2 - name9 - 109-103

I would like to request to view the data in the following order:

grade - score - name - record_id - parentid
----------------------------------------------------------
1 - 1 - name1 - 101-0
2 - 1 - name4 - 104-101
2 - 2 - name5 - 105-101
1 - 1 - name2 - 102-0
2 - 1 - name6 - 106-102
2 - 2 - Marque7 - 107-102
1 - 1 - name3 - 103-0
2 - 1 - name8 - 108-103
2 - 2 - name9 - 109-103

The simple query to get this sequence of result set is adequate. No need to fill or other characteristics.

Thank you.

All the columns in your sample have same agenda, so it is not possible to tell which column you want to sisters of the order. I'll assume by its name:

select  *
  from  your_table
  start with parentid = 0
  connect by parentid = prior record_id
  order siblings by name
/

For example:

with sample_table as (
                      select 1 grade,1 score,'name1' name,101 record_id,0 parentid from dual union all
                      select 1,2,'name2',102,0 from dual union all
                      select 1,3,'name3',103,0 from dual union all
                      select 2,1,'name4',104,101 from dual union all
                      select 2,2,'name5',105,101 from dual union all
                      select 2,1,'name6',106,102 from dual union all
                      select 2,2,'name7',107,102 from dual union all
                      select 2,1,'name8',108,103 from dual union all
                      select 2,2,'name9',109,103 from dual
                     )
select  *
  from  sample_table
  start with parentid = 0
  connect by parentid = prior record_id
  order siblings by name
/

     GRADE      SCORE NAME   RECORD_ID   PARENTID
---------- ---------- ----- ---------- ----------
         1          1 name1        101          0
         2          1 name4        104        101
         2          2 name5        105        101
         1          2 name2        102          0
         2          1 name6        106        102
         2          2 name7        107        102
         1          3 name3        103          0
         2          1 name8        108        103
         2          2 name9        109        103

9 rows selected.

SQL>

SY.

Tags: Database

Similar Questions

hierarchical sql (part 2) - error: duplicate rows
Following up on my previous question:

This hierarchical query has the duplicate in the result set rows:

Select record_id, parentid, name from my_table
connect by parentid = prior record_id
siblings arrested by name;

Thanks for help.

----------

My table has five columns:

rank (number)
score (number)
name (varchar)
record_id (number)
ParentId (number)

A parentid is a record_id at a higher level in the hierarchy.
A parentid is a record_id at a higher level in the hierarchy.
Rank and Score are used to control (tri) each level of the hierarchy.

The data looks like this:

grade - score - name - record_id - parentid

--------------------------------------------------------------------------------
1 - 1 - name1 - 101-0
1 - 2 - name2 - 102-0
1 - 3 - name3 - 103-0
2 - 1 - name4 - 104-101
2 - 2 - name5 - 105-101
2 - 1 - name6 - 106-102
2 - 2 - Marque7 - 107-102
2 - 1 - name8 - 108-103
2 - 2 - name9 - 109-103

Desired results set:

grade - score - name - record_id - parentid
--------------------------------------------------------------------------------
1 - 1 - name1 - 101-0
2 - 1 - name4 - 104-101
2 - 2 - name5 - 105-101
1 - 1 - name2 - 102-0
2 - 1 - name6 - 106-102
2 - 2 - Marque7 - 107-102
1 - 1 - name3 - 103-0
2 - 1 - name8 - 108-103
2 - 2 - name9 - 109-103
Missing the BEGINNING WITH...?
```
SQL> select record_id, parentid, name
  2  from my_table
  3  --Added START WITH
  4  start with parentid = 0
  5  connect by parentid = prior record_id
  6  order siblings by name;

 RECORD_ID   PARENTID NAME
---------- ---------- ----------
       101          0 name1
       104        101 name4
       105        101 name5
       102          0 name2
       106        102 name6
       107        102 name7
       103          0 name3
       108        103 name8
       109        103 name9

9 rows selected.
```

Recursive, HIERARCHICAL query problem

Hello!

I have a hierarchical SQL help 'connect' with the result in the following dataset (simplified down point):

ID       | Parent |   IsData 
----------------------------
A        |        |    N     
  A1     |   A    |    N     
  A2     |   A    |    N     
    X1   |   A2   |    Y     
B        |        |    N     
  B1     |   B    |    N     
    X2   |   B1   |    Y     
  B2     |   B    |    N     
C        |        |    N     
  C1     |   C    |    N

There are entries of data (X 1 and X 2) and structural (year, Bn, Cn) entries. Some nodes in the tree structure have entries of data like childs, some do not.

What I need is a third column that says "this node is a node of data below recursively' as indicated below:

 

ID       | Parent |   IsData    |  HasRecursiveDataChilds
----------------------------------------------------------
A        |        |     N       |        Y
  A1     |   A    |     N       |        N
  A2     |   A    |     N       |        Y
    X1   |   A2   |     Y       |        Y/N (doesn't matter)
B        |        |     N       |        Y
  B1     |   B    |     N       |        Y
    X2   |   B1   |     Y       |        Y/N (doesn't matter)
  B2     |   B    |     N       |        N
C        |        |     N       |        N
  C1     |   C    |     N       |        N

I currently do on the client side, after that I got all of the data, which involves an iteration of the full dataset among at least once (if you are familiar with the data). Data is accessible from another (web) application, I want to spend the logic to the server.

I wonder if this can be done with a simple SQL statement without using the PL/SQL to iterate over the game "manually" of results. I can use all of the features of 10g. Performance is not quite important, so wild sub-sub sub-sub-sub-queries would be ok.

Thanks for your help,
Marcus

OK, this is somewhat a fudge...

SQL> ed
Wrote file afiedt.buf

  1  with t as (select 'A' as ID, null as parent, 'N' as IsData from dual union all
  2             select 'A1','A','N' from dual union all
  3             select 'A2','A','N' from dual union all
  4             select 'X1','A2','Y' from dual union all
  5             select 'B', null, 'N' from dual union all
  6             select 'B1','B','N' from dual union all
  7             select 'X2','B1','Y' from dual union all
  8             select 'B2','B','N' from dual union all
  9             select 'C',null,'N' from dual union all
 10             select 'C1','C','N' from dual)
 11  -- END OF TEST DATA
 12  select lpad(' ',(level-1)*2,' ')||id as id, parent, isdata
 13        ,NVL((select max(t2.isdata)
 14          from   t t2
 15          connect by parent = prior id
 16          start with parent = t.id),'N') as recursivechilddata
 17  from t
 18  connect by parent = prior id
 19* start with parent is null
SQL> /

ID         PARENT     ISDATA     RECURSIVECHILDDATA
---------- ---------- ---------- -------------------------
A                     N          Y
  A1       A          N          N
  A2       A          N          Y
    X1     A2         Y          N
B                     N          Y
  B1       B          N          Y
    X2     B1         Y          N
  B2       B          N          N
C                     N          N
  C1       C          N          N

10 rows selected.

SQL>

Published by: BluShadow on April 9, 2009 10:59
removed the redundant code

recursion in pl/sql
Hi all
I need to know all the nested procedure or the list of functions that are called by a function or a special procedure. I am trying to write a code that I use recursion. I am facing problem to store values in the clause of the SELECT INTO statement because multiple procs may not be stored on a single clue that is to say proclist_v (i). Should I use another type of another object that pl/sql table.please give me an idea

CREATE or REPLACE PACKAGE finddblink_pack
IS
TYPE NameType IS the TABLE OF
dba_dependencies.referenced_name%type
INDEX OF DIRECTORY;
PROCEDURE finddblink_proc (proc_par NameType)
END finddblink_pack;
/

CREATE or REPLACE PACKAGE finddblink_pack BODY
IS

PROCEDURE finddblink_proc (proc_par NameType)
IS
proclist_v NameType;
BEGIN
BECAUSE me in func_par. FIRST... func_par. LAST LOOP
SELECT referenced_name
IN proclist_v (i)
OF dba_dependencies
WHERE name = upper (func_par (i))
AND referenced_type ('FUNCTION', 'PROCEDURE', 'BODY of PACKAGE');
finddblink (proclist_v);
END LOOP;

END finddblink_proc;
END finddblink_pack;
Well, you don't need recursive plsql for this. Hierarchical SQL will do:
```
SQL> CREATE OR REPLACE
  2    PROCEDURE P1
  3      IS
  4      BEGIN
  5          NULL;
  6  END;
  7  /

Procedure created.

SQL> CREATE OR REPLACE
  2    PROCEDURE P2
  3      IS
  4      BEGIN
  5          P1;
  6  END;
  7  /

Procedure created.

SQL> CREATE OR REPLACE
  2    PROCEDURE P3
  3      IS
  4      BEGIN
  5          P2;
  6  END;
  7  /

Procedure created.

SQL> SELECT  referenced_name
  2    FROM  dba_dependencies
  3    START WITH name = 'P3'
  4    CONNECT BY prior referenced_name = name
  5    AND referenced_type in('FUNCTION','PROCEDURE','PACKAGE BODY')
  6  /

REFERENCED_NAME
----------------------------------------------------------------
P2
P1

SQL> 
```
SY.

By using a WHERE clause in the tree APEX

Hi all -
I have a hierarchical SQL query that I display as a tree of the APEX.
Here is my sample application:
https://Apex.Oracle.com/pls/Apex/f?p=32581:29
Login: guest
PW: app_1000
workspace: leppard
I try to add a WHERE clause so that only nodes with the lowest-level children are displayed, i.e. something like "WHERE connect_by_isleaf = 0 OR level = 5'"
The tree of query with where clause works very well in the SQL command window, but when I add the WHERE clause to my tree apex page no longer displays anything. Is this a problem with APEX or y at - it another way to filter my results?
Thanks in advance for your suggestions,
John

Connect occurs in the first place, the where the provision is applied to these results, effectively cutting in the hierarchical structure. Since the apex has build a hierarchical structure of the query, it relies on the virtual level, which is butchered by applying the where clause. It's a miracle that you're not even receive errors because I expect almost an incorrect json array was built. With no top-level from and only level 5 or end nodes, there is no structure to present: the two apex cannot set a correct representation and not jstree neither. In your request, you will see your "root nodes", but it is not representable.

I don't know why you want to present it in a tree? Levels both connect_by_is_leaf = 5 will give you everything just a list of nodes without any hierarchical structure.

The best thing to do is to use a subquery to limit first your dataset and then use it for the basic tree query, this way you violate one of these vital columns.

For example if you want only nodes and their immediate parent, you can opt for something like this (fast on some testdata model):
```
with dataset as (
select node_id, parent_id
   from treedata
  where connect_by_isleaf = 0
connect by prior node_id = parent_id
  start with parent_id = 0
),
dataset2 as (
select node_id, parent_id
  from dataset
union all
select node_id, null parent_id
  from treedata
where node_id in (select parent_id from dataset)
)
select level, node_id
   from dataset2
connect by prior node_id = parent_id
  start with parent_id is null
```

A hierarchical report query SQL formatting to display the output as a tree

4.2.1
THM:2
Hello world
I have a simple request which when I run it should show in a tree like structure as seen in this link
Hierarchical queries in Oracle SQL
Any ideas on how I can get this output show upward in an area of report page apex? I tried to put it in a classic report, and it shows just all in one line given the length and spacing. I don't want to use the tree for this.
Thank you!

Ryansun-Oracle wrote:

I have a simple request which when I run it should show in a tree like structure as seen in this link

Hierarchical queries in Oracle SQL

Any ideas on how I can get this output show upward in an area of report page apex? I tried to put it in a classic report, and it shows just all in one line given the length and spacing. I don't want to use the tree for this.

I found that the standard reports and report templates are not really useful in this situation.

According to exactly how it should be used, you might want to watch a dynamic list using one of the models of hierarchical list, but I tend to use this Tyler Muth & Rob Van Wijk. This can be used in a region in connection with a column named custom report template.

How to order a tree balanced with SQL hierarchical queries
by searching the forum I found this

Re: Generate tree balanced with SQL hierarchical queries

is there a way I can order this tree in alphabetical order so that the result looks like

LEVEL INDENTED_ENAME
---------- --------------------
1 KING BED
2 BLAKE
3 ALLEN
3 JAMES
MARTIN 3
3 TURNER
WARD 3
2 CLARK
3 MILLER
2 JONES
3 FORD
4 SMITH
3 SCOTT
4 ADAMS

-the original query-

SELECT THE LEVEL
, LPAD (' ', 3 * LEVEL) | Ename AS indented_ename
FROM scott.emp
START WITH mgr IS NULL
CONNECT BY PRIOR empno = mgr
;

LEVEL INDENTED_ENAME
---------- --------------------
1 KING BED
2 JONES
3 SCOTT
4 ADAMS
3 FORD
4 SMITH
2 BLAKE
3 ALLEN
WARD 3
MARTIN 3
3 TURNER
3 JAMES
2 CLARK
3 MILLER
Hello

Bodin wrote:
Hi Frank, I can order it selectively depending on the level, which means that only siblings stopped at third level, but rest of the brothers and sisters remain Nations United ordered

It's actually quite difficult. You can "ORDER of brothers and SŒURS BY CASE... ', like this:
```
SELECT  LEVEL
,      LPAD (' ', 3 * LEVEL) || ename     AS indented_ename
FROM      scott.emp
START WITH        mgr     IS NULL
CONNECT BY         mgr      = PRIOR empno
ORDER SIBLINGS BY  CASE
               WHEN  job = 'MANAGER'  THEN  ename
                                          ELSE  NULL
             END
;
```
In this case to get desired results in table scott.emp, as the lines are LEVEL = 2 if and only if use = "MANAGER".
But if you reference LEVEL in the CASE expression (for example, if you replace ' job = 'MANAGER' ' with "2 LEVEL =" above "), then you will get the error" ORA-00976: LEVEL, PRIOR or ROWNUM not allowed here. "
The best way I can think to do exactly what you asked is to do 2 CONNECT BY queries; one just to get the LEVEL and the other for the brothers and SŒURS ORDER BY:
{code}
WITH got_lvl AS
(
SELECT LEVEL AS lvl
Bishop
empno
ename
FROM scott.emp
START WITH mgr IS NULL
CONNECT BY PRIOR empno = mgr
)
SELECT lvl
, LPAD (' ', 3 * LEVEL) | Ename AS indented_ename
OF got_lvl
START WITH mgr IS NULL
CONNECT BY PRIOR empno = mgr
BROTHERS AND SŒURS OF ORDER OF CASES
ONCE lvl = 2 THEN ename
ANOTHER NULL
END
;
{code}
Why you can't simply "Brothers and SŒURS of ORDER BY ename" at all levels? If all you care is the order of the items of LEVEL = 2, then this is probably the most effective and simplest way. It really hurt anything if nodes on levels 3, 4, 5,... are in order, too?

Here's something you can do if you want to order by different unique things to different levels:
{code}
WITH got_sort_key AS
(
SELECT LEVEL AS lvl
, LPAD (' ', 3 * LEVEL) | Ename AS indented_ename
empno
SYS_CONNECT_BY_PATH (LPAD (CASE
WHEN LEVEL = 2
THEN ename
Of OTHER TO_CHAR (empno)
END
10
)
, ','
) AS sort_key
FROM scott.emp
START WITH mgr IS NULL
CONNECT BY PRIOR empno = mgr
)
SELECT lvl
indented_ename
empno
OF got_sort_key
ORDER BY sort_key
;
{code}
However, all possible values of the CASE expression must uniquely identify the node; otherwise, the output won't necessarily hierarchical order. You can assign arbitrary unique IDS to the lines (using the ROW_NUMBER analytic function, for example), but that requires another subquery and is also complex and perhaps as ineffective as the solution above with 2 garages CONNECT.

Building the tree balanced with SQL hierarchical queries

Hi all

I have the following hierarchical data with different levels of subtree:

A0
-A001
-A00101
A1
-A101
A2
-A201
-A20101
-A201010001

A0 subtree has 3 levels, A1 subtree has 2 levels and subtree of the A3 is level 4. I want to generate a tree balanced on the data with all levels of the subtree equal to the maximum number of levels available in the whole tree which, in this particular case, is 4.

I don't know that it is possible with SQL. Script to generate the above mentioned are as below:

CREATE TABLE codes_tree
(node_id VARCHAR2 (10))
parent_node_id VARCHAR2 (10)
);
INSERT INTO codes_tree VALUES ('A0', NULL);
INSERT INTO codes_tree VALUES ('A001', 'A0');
INSERT INTO codes_tree VALUES ('A00101', 'A001');
---
INSERT INTO codes_tree VALUES ('A1', NULL);
INSERT INTO codes_tree VALUES ('A101', 'A1');
---
INSERT INTO codes_tree VALUES ('A2', NULL);
INSERT INTO codes_tree VALUES ('A201', 'A2');
INSERT INTO codes_tree VALUES ('A20101', 'A201');
INSERT INTO codes_tree VALUES ('A201010001', 'A20101');

Any help will be much appreciated.

Thank you... Best regards

Published by: naive2Oracle on May 12, 2011 19:40

Published by: naive2Oracle on May 12, 2011 19:41

Hello

Of course, you can do it in SQL.
One way is to take the normal output of hierarchical and manipulate the result set so that the leaves are repeated as often as necessary to make all branches of the same length. I have Oracle 10.2 available right now, so here's a solution that will work in Oracle 10 (and more):

WITH     original_hierarchy     AS
(
     SELECT     node_id
     ,     LEVEL               AS lvl
     ,     CONNECT_BY_ISLEAF     AS isleaf
     ,     ROWNUM               AS rnum
     FROM     codes_tree
     START WITH     parent_node_id     IS NULL
     CONNECT BY     parent_node_id     = PRIOR node_id
)
,     got_max_lvl     AS
(
     SELECT     o.*
     ,     MAX (lvl) OVER ()     AS max_lvl
     FROM     original_hierarchy     o
)
SELECT       LPAD ( ' '
            , 3 * ( ( d.lvl
                 + NVL (c.rnum, 1)
                 - 1
                 )
               - 1
               )
            ) || CASE
               WHEN c.rnum > 1
               THEN '*' || d.node_id || '*'
               ELSE        d.node_id
              END          AS display_id
FROM            got_max_lvl     d
LEFT OUTER JOIN       got_max_lvl     c  ON     d.isleaf     = 1
                       AND     c.rnum          <= 1 + d.max_lvl - d.lvl
ORDER BY  d.rnum
,       c.rnum
;

With the help of Oracle 11.2, it would be preferable to generate original_hierarchy as above, but to manipulate using a WITH recursive clause.
Analytical functions often interfere with CONNECT BY, so I used a separate subquery to get max_lvl, do CONNECT BY in a sub-querry and analytic function in a separate subquery. I don't know what is needed on all versions.

Output of your sample data:

DISPLAY_ID
-------------------------------
A0
   A001
      A00101
         *A00101*
A1
   A101
      *A101*
         *A101*
A2
   A201
      A20101
         A201010001

Below is a generic version of the same query, which I used to test this on scott.emp:

DEFINE     table_name     = scott.emp
DEFINE     id_col          = empno
DEFINE     parent_id_col     = mgr
DEFINE     display_col     = ename

WITH     original_hierarchy     AS
(
     SELECT     &display_col          AS display_txt
     ,     LEVEL               AS lvl
     ,     CONNECT_BY_ISLEAF     AS isleaf
     ,     ROWNUM               AS rnum
     FROM     &table_name
     START WITH     &parent_id_col     IS NULL
     CONNECT BY     &parent_id_col     = PRIOR &id_col
)
,     got_max_lvl     AS
(
     SELECT     o.*
     ,     MAX (lvl) OVER ()     AS max_lvl
     FROM     original_hierarchy     o
)
SELECT       LPAD ( ' '
            , 3 * ( ( d.lvl
                 + NVL (c.rnum, 1)
                 - 1
                 )
               - 1
               )
            ) || CASE
               WHEN c.rnum > 1
               THEN '*' || d.display_txt || '*'
               ELSE        d.display_txt
              END          AS display_id
FROM            got_max_lvl     d
LEFT OUTER JOIN       got_max_lvl     c  ON     d.isleaf     = 1
                       AND     c.rnum          <= 1 + d.max_lvl - d.lvl
ORDER BY  d.rnum
,       c.rnum
;

Output:

DISPLAY_ID
-----------------------------
KING
   JONES
      SCOTT
         ADAMS
      FORD
         SMITH
   BLAKE
      ALLEN
         *ALLEN*
      WARD
         *WARD*
      MARTIN
         *MARTIN*
      TURNER
         *TURNER*
      JAMES
         *JAMES*
   CLARK
      MILLER
         *MILLER*

Published by: Frank Kulash, May 13, 2011 06:38
Adding the generic version

SQL SELECT to hierarchical tables: START WITH... CONNECT BY...
This seems to be a simple problem, but I think that I have enough intelligence SQL to solve.

I have a table called DE_DOMAIN. The columns of interest are:
DOMAIN_ID - PK
NAME
PARENT_ID - FK (can be NULL), poining to CF of the parent

What I want is: Returns a hierarchical list, containing: column 3 above as well as the NAME of the parent.

Regardless of the name of the parent, it works fine:
SELECT DOMAIN_ID, PARENT_ID, level
OF DE_DOMAIN
WHERE SUPERDOMAINE = 2673
Start by PARENT_ID IS NULL
Connect DOMAIN_ID PARENT_ID = prior
BROTHERS AND SŒURS ORDER BY NAME ASC

and I get:
11 rec_11 1 Null
15 rec_15 1 Null
16 1 Null rec_16
17 1 Null rec_17
22 17 2 rec_22
1 2 17 rec_1
rec_25 25 17 2
2 2 17 rec_2

i.e. records with PK = 1, 22, 25, 2 have all like parent record with PK = 17, then the new column name, they must bear the name of the parent (i.e. rec_17).

A simple idea?
Thank you very much.
Hello

You can use the FIRST operator in the SELECT clause.
I don't have a version of your table, so I'll use scott.emp to illustrate:
```
SELECT     empno
,     ename
,     mgr
,     PRIOR ename    AS mgr_name
FROM     scott.emp
START WITH     mgr     IS NULL
CONNECT BY     mgr     = PRIOR empno
ORDER SIBLINGS BY     ename
;
```
Output:
```
.    EMPNO ENAME             MGR MGR_NAME
---------- ---------- ---------- ----------
      7839 KING
      7698 BLAKE            7839 KING
      7499 ALLEN            7698 BLAKE
      7900 JAMES            7698 BLAKE
      7654 MARTIN           7698 BLAKE
      7844 TURNER           7698 BLAKE
      7521 WARD             7698 BLAKE
      7782 CLARK            7839 KING
      7934 MILLER           7782 CLARK
      7566 JONES            7839 KING
      7902 FORD             7566 JONES
      7369 SMITH            7902 FORD
      7788 SCOTT            7566 JONES
      7876 ADAMS            7788 SCOTT
```

Problem (hierarchical data) resolved by using SQL

Hi I have hierarchical data stored in a table with 3 columns
Item1, item2, relationship

relationship column defines if the element1 is parent of item2 or item1 item2 Chile

I have a fourth column in the same table indicating the entire hierarchy of the family separated by colons;

example: a1, a2, child then output will be a1:a2.

I have tried a lot of things to do with a sql but failed and finally did using the recursive function

Hoping someone can solve the problem in sql

You will find examples of data to work with as well as the desired result:

select 'p1' as element1, 'p2' as element2, 'parent' as relationship, 'p1' as output  from dual
union all
select 'p2', 'p1', 'child', 'p1:p2' from dual
union all
select 'p1', 'p3', 'parent', 'p1' from dual
union all
select 'p3', 'p1', 'child', 'p1:p3' from dual
union all
select 'p2', 'p4', 'parent', 'p1:p2' from dual
union all
select 'p4', 'p2', 'child', 'p1:p2:p4' from dual
union all
select 'p2', 'p5', 'parent', 'p1:p2' from dual
union all
select 'p5', 'p2', 'child', 'p1:p2:p5' from dual
union all
select 'p3', 'p6', 'parent', 'p1:p3' from dual
union all
select 'p6', 'p3', 'child', 'p1:p3:p6' from dual
union all
select 'p5', 'p7', 'parent', 'p1:p2:p5' from dual
union all
select 'p7', 'p5', 'child', 'p1:p2:p5:p7' from dual
union all
select 'p5', 'p8', 'parent', 'p1:p2:p5' from dual
union all
select 'p8', 'p5', 'child', 'p1:p2:p5:p8' from dual
union all
select 'b1', 'b2', 'parent', 'b1' from dual
union all
select 'b2', 'b1', 'child', 'b1:b2' from dual
union all
select 'b2', 'b3', 'parent', 'b1:b2' from dual
union all
select 'b3', 'b2', 'child', 'b1:b2:b3' from dual
union all
select 'c1', 'c2', 'parent', 'c1' from dual
union all
select 'c2', 'c1', 'child', 'c1:c2' from dual
union all
select 'c1', 'c3', 'parent', 'c1' from dual
union all
select 'c3', 'c1', 'child', 'c1:c3' from dual
union all
select 'c3', 'c5', 'parent', 'c1:c3' from dual
union all
select 'c5', 'c3', 'child', 'c1:c3:c5' from dual
union all
select 'c3', 'c6', 'parent', 'c1:c3' from dual
union all
select 'c6', 'c3', 'child', 'c1:c3:c6' from dual
union all
select 'c3', 'c7', 'parent', 'c1:c3' from dual
union all
select 'c7', 'c3', 'child', 'c1:c3:c7' from dual
union all
select 'a1', null, 'parent', 'a1' from dual

Kind regards
Amit

don't worry, I fixed it...

with t as(
select 'p1' as element1, 'p2' as element2, 'parent' as relationship  from dual
union all
select 'p2', 'p1', 'child' from dual
union all
select 'p1', 'p3', 'parent'from dual
union all
select 'p3', 'p1', 'child' from dual
union all
select 'p2', 'p4', 'parent' from dual
union all
select 'p4', 'p2', 'child' from dual
union all
select 'p2', 'p5', 'parent' from dual
union all
select 'p5', 'p2', 'child' from dual
union all
select 'p3', 'p6', 'parent' from dual
union all
select 'p6', 'p3', 'child' from dual
union all
select 'p5', 'p7', 'parent' from dual
union all
select 'p7', 'p5', 'child' from dual
union all
select 'p5', 'p8', 'parent' from dual
union all
select 'p8', 'p5', 'child' from dual
union all
select 'b1', 'b2', 'parent' from dual
union all
select 'b2', 'b1', 'child' from dual
union all
select 'b2', 'b3', 'parent' from dual
union all
select 'b3', 'b2', 'child' from dual
union all
select 'c1', 'c2', 'parent' from dual
union all
select 'c2', 'c1', 'child' from dual
union all
select 'c1', 'c3', 'parent' from dual
union all
select 'c3', 'c1', 'child' from dual
union all
select 'c3', 'c5', 'parent' from dual
union all
select 'c5', 'c3', 'child' from dual
union all
select 'c3', 'c6', 'parent' from dual
union all
select 'c6', 'c3', 'child' from dual
union all
select 'c3', 'c7', 'parent' from dual
union all
select 'c7', 'c3', 'child' from dual
union all
select 'a1', null, 'parent' from dual
), p as (
select * from t where relationship='parent')
,q as (
select * from t where relationship='child'
)
select element1,element2,relationship,path from (select path,element1,'parent' relationship,element2 from (
select path,element1,element2,row_number() over (partition by element1,element2 order by length(path) desc) rnum from (
select ltrim(sys_connect_by_path(element1,':'),':') path,element1,element2 from p connect by prior element2=element1
)) where rnum=1
union all
select path,element1,'child' relationship,element2 from (
select path,element1,element2,row_number() over (partition by element1,element2 order by length(path) desc) rnum from (
select ltrim(connect_by_root element2||sys_connect_by_path(element1,':'),':') path,element1,element2 from q connect by prior element1=element2
)) where rnum=1)

Ravi Kumar

How to convert the hierarchical query of SQL Server (CTE) to Oracle?
How to convert the hierarchical query of SQL Server (CTE) to Oracle?

WITH cte (col1, col2) AS
(
SELECT col1, col2
FROM dbo. [tb1]
WHERE col1 = 12
UNION ALL
SELECT c.col1, c.col2
FROM dbo. [tb1] AS c INNER JOIN cte AS p ON c.col2 = p.col1
)
DELETE one
FROM dbo. [tb1] AS an INNER JOIN b cte
ON a.col1 = b.col1
Hello
Something like this maybe:
```
DELETE FROM dbo.tb1 a
 WHERE EXISTS (
  SELECT 1
    FROM dbo.tb1 b
  WHERE a.co11 = b.col1
      AND a.col2 = b.col2
   START WITH b.col1 = 12
  CONNECT BY b.col2 = PRIOR b.col1)
```
Although you need to do here is to check that CONNECT it BY SELECT, returns records you wait first, then the DELETION should work too.

SQL Query hierarchical select 2 level parent over sheet.

Here, we use Oracle 11 g R1. Here is a sample of our Tables of the employee.
I want you select all managers managers above them (which we call the Site Manager) and direct (that we call the Service Manager)
Here is the example for the base table.
```
SET DEFINE OFF;
CREATE TABLE EMP_SAMPLE 
(
  AGENT_ID VARCHAR2(100 BYTE) 
, FULL_NAME VARCHAR2(100 BYTE) 
, AGENT_MANAGER VARCHAR2(100 BYTE) 
) ;

Insert into EMP_SAMPLE (AGENT_ID,FULL_NAME,AGENT_MANAGER) values ('JS001','JOHN SMITH',null);
Insert into EMP_SAMPLE (AGENT_ID,FULL_NAME,AGENT_MANAGER) values ('AL001','ANN LEE','JS001');
Insert into EMP_SAMPLE (AGENT_ID,FULL_NAME,AGENT_MANAGER) values ('JD001','JOHN DOE','AL001');
Insert into EMP_SAMPLE (AGENT_ID,FULL_NAME,AGENT_MANAGER) values ('MB002','MARY BAKER','AL001');
Insert into EMP_SAMPLE (AGENT_ID,FULL_NAME,AGENT_MANAGER) values ('HM003','HOWARD MONROE','MB002');
Insert into EMP_SAMPLE (AGENT_ID,FULL_NAME,AGENT_MANAGER) values ('RM001','ROBYN MILLER','MB002');
Insert into EMP_SAMPLE (AGENT_ID,FULL_NAME,AGENT_MANAGER) values ('DJ002','DAVID JONES','RM001');
Insert into EMP_SAMPLE (AGENT_ID,FULL_NAME,AGENT_MANAGER) values ('WW001','WENDY WONG','MB002');
Insert into EMP_SAMPLE (AGENT_ID,FULL_NAME,AGENT_MANAGER) values ('PB001','PETER RABBIT','JS001');
Insert into EMP_SAMPLE (AGENT_ID,FULL_NAME,AGENT_MANAGER) values ('BB002','BEN BUNNY','PB001');
Insert into EMP_SAMPLE (AGENT_ID,FULL_NAME,AGENT_MANAGER) values ('TM001','TONY MILLER','BB002');
Insert into EMP_SAMPLE (AGENT_ID,FULL_NAME,AGENT_MANAGER) values ('PP002','PETER PARKER','RM001');
Insert into EMP_SAMPLE (AGENT_ID,FULL_NAME,AGENT_MANAGER) values ('PP003','PEPPA PIG','PB001');
Insert into EMP_SAMPLE (AGENT_ID,FULL_NAME,AGENT_MANAGER) values ('DB002','daniel baker','HM003');
Insert into EMP_SAMPLE (AGENT_ID,FULL_NAME,AGENT_MANAGER) values ('TL001','Tom Lee','WW001');
Insert into EMP_SAMPLE (AGENT_ID,FULL_NAME,AGENT_MANAGER) values ('MS001','Mary Smith',null);
```
With the example like this data, I would like to that the result looks similar to
Name of the Manager
MARY BAKER - Manager of the Site
HOWARD MONROE - Service Manager
ROBYN MILLER - Service Manager
WENDY WOND - Service Manager
PETER RABBIT Site Manager
BEN BUNNY - Service Manager
I guess I should use the hierarchical query to achieve this. I googled to see all the solutions, I tried this under request
```
SELECT agent_id, agent_manager, RPAD('.', (level-1)*2, '.') || full_name AS tree,
       level, CONNECT_BY_ROOT agent_id as root_id
       ,CONNECT_BY_ISLEAF AS is_leaf

FROM EMP_SAMPLE
START WITH agent_manager IS NULL
CONNECT BY agent_manager = PRIOR agent_id
ORDER SIBLINGS BY agent_id;
```
But it seems that the level always starts from the root at the top of the page. I wonder, is it possible to identify even just the level two above the sheet.
Here is some basic information. We are working on our oracle APEX application. One of the reports can be filtered by the Manager. The list currently shows a little all managers, regardless of what they are 1 senior as general manager, or the first line as a Service Manager Manager. Now, users want to be able to select only Service Manager and Site Manager.
There is a DB work for updating the table Employee of LDAP. And the LDAP structure is not tidy this (we have some staff members who have no Manager, they are not even CEO). The Administrator told us that it is too busy to store it.
We have about 20,000 employees in total including 800 East of managers.
Thanks in advance.
Ann

Hi, Ann.

Ann586341 wrote:

Here, we use Oracle 11 g R1. Here is a sample of our Tables of the employee.

I want you select all managers managers above them (which we call the Site Manager) and direct (that we call the Service Manager)

Here is the example for the base table.
1. TOGETHER TO DEFINE
2. CREATE TABLE EMP_SAMPLE
3. (
4. AGENT_ID VARCHAR2 (100 BYTE)
5. FULL_NAME VARCHAR2 (100 BYTE)
6. AGENT_MANAGER VARCHAR2 (100 BYTE)
7. ) ;
8. Insert into EMP_SAMPLE (AGENT_ID, FULL_NAME, AGENT_MANAGER) values ('JS001', 'JOHN SMITH', null);
9. Insert into EMP_SAMPLE (AGENT_ID, FULL_NAME, AGENT_MANAGER) values ('AL001', "ANN LEE", "JS001");
10. Insert into EMP_SAMPLE (AGENT_ID, FULL_NAME, AGENT_MANAGER) values ('JD001', 'JOHN DOE', "AL001");
11. Insert into EMP_SAMPLE (AGENT_ID, FULL_NAME, AGENT_MANAGER) values ('MB002', "MARY BAKER", "AL001");
12. Insert into EMP_SAMPLE (AGENT_ID, FULL_NAME, AGENT_MANAGER) values ('HM003', 'HOWARD MONROE', 'MB002');
13. Insert into EMP_SAMPLE (AGENT_ID, FULL_NAME, AGENT_MANAGER) values ('RM001", 'ROBYN MILLER', 'MB002');
14. Insert into EMP_SAMPLE (AGENT_ID, FULL_NAME, AGENT_MANAGER) values ('DJ002', 'DAVID JONES', "RM001");
15. Insert into EMP_SAMPLE (AGENT_ID, FULL_NAME, AGENT_MANAGER) values ('WW001', "WENDY WONG", "MB002");
16. Insert into EMP_SAMPLE (AGENT_ID, FULL_NAME, AGENT_MANAGER) values ('PB001', 'PETER RABBIT', "JS001");
17. Insert into EMP_SAMPLE (AGENT_ID, FULL_NAME, AGENT_MANAGER) values ('BB002', "BEN BUNNY", "PB001");
18. Insert into EMP_SAMPLE (AGENT_ID, FULL_NAME, AGENT_MANAGER) values ('TM001', 'TONY MILLER', "BB002");
19. Insert into EMP_SAMPLE (AGENT_ID, FULL_NAME, AGENT_MANAGER) values ('PP002","PETER PARKER","RM001");
20. Insert into EMP_SAMPLE (AGENT_ID, FULL_NAME, AGENT_MANAGER) values ('PP003', 'PEPPA PIG', "PB001");
21. Insert into EMP_SAMPLE (AGENT_ID, FULL_NAME, AGENT_MANAGER) values ('DB002', "daniel baker", "HM003");
22. Insert into EMP_SAMPLE (AGENT_ID, FULL_NAME, AGENT_MANAGER) values ('TL001', "Tom Lee", "WW001");
23. Insert into EMP_SAMPLE (AGENT_ID, FULL_NAME, AGENT_MANAGER) values ('MS001', 'Mary Smith', null);
With the example like this data, I would like to that the result looks similar to

Name of the Manager

MARY BAKER - Manager of the Site

HOWARD MONROE - Service Manager

ROBYN MILLER - Service Manager

WENDY WOND - Service Manager

PETER RABBIT Site Manager

BEN BUNNY - Service Manager

I guess I should use the hierarchical query to achieve this. I googled to see all the solutions, I tried this under request
1. SELECT agent_id, agent_manager, RPAD ('.) (', (level 1) * 2, '.') || full_name LIKE tree,
2. level, agent_id CONNECT_BY_ROOT as root_id
3. CONNECT_BY_ISLEAF AS is_leaf
4. OF EMP_SAMPLE
5. START WITH agent_manager IS NULL
6. CONNECT BY PRIOR agent_id = agent_manager
7. Brothers and SŒURS of ORDER BY agent_id;
But it seems that the level always starts from the root at the top of the page. I wonder, is it possible to identify even just the level two above the sheet.

Here is some basic information. We are working on our oracle APEX application. One of the reports can be filtered by the Manager. The list currently shows a little all managers, regardless of what they are 1 senior as general manager, or the first line as a Service Manager Manager. Now, users want to be able to select only Service Manager and Site Manager.

There is a DB work for updating the table Employee of LDAP. And the LDAP structure is not tidy this (we have some staff members who have no Manager, they are not even CEO). The Administrator told us that it is too busy to store it.

We have about 20,000 employees in total including 800 East of managers.

Thanks in advance.

Ann

Here's one way:

WITH leaves LIKE

(

SELECT agent_id

Of emp_sample

LESS

SELECT agent_manager

Of emp_sample

)

got_lvl AS

(

SELECT-full_name agent_id

LEVEL AS lvl

Of emp_sample

START WITH agent_id IN)

SELECT agent_id

Sheets

)

CONNECTION BY agent_id = agent_manager PRIOR

)

SELECT full_name

CASE

WHEN MAX (lvl) = 3

THEN 'Site Manager'

ANOTHER "Service Manager"

END AS title

OF got_lvl

GROUP BY full_name-, agent_id

WITH MIN (lvl) > = 2

AND MAX (lvl)<=>

ORDER BY full_name

;

Output:

FULL_NAME TITLE

-------------------- ---------------

BEN BUNNY Service Manager

MONROE HOWARD Service Manager

MARY BAKER Site Manager

PETER RABBIT Site Manager

ROBYN MILLER Service Manager

WENDY WONG Service Manager

Thanks for posting the CREATE TABLE and INSERT statements; It's very useful!

LEVEL does not always begin with the root; It starts with the lines that meet the condition to START WITH. If you have a START WITH condition that only roots meet (as in the query you posted) then, Yes, LEVEL will start with the roots. If you have a condition START WITH answering the criteria only the leaves (as in the query I posted), then LEVEL will start with the leaves.

I guess just some of your needs. If seems that a 'Site Manager' is the grandparent of a leaf, but "John Smith" (which is the grandmother of "John Doe", a sheet) is not considered to be a 'Site Manager'. Similarly, it seems that a 'Service Manager' is the parent (but not the grand-parent) of the leaf, but "Ann Lee" (the mother of "John Doe") is not considered to be a 'Service Manager' for a reason any.

I guess that full_name is unique. If full_name is not unique, but agent_id is, then you will need a comment a few line above ends. (I guess that agent_id is unique and not NULL).

Problem with hierarchical query in function PL\SQL

I have a simple table containing the ID of the parent

-Create table

create the table1 table:

(

ID NUMBER (12) not null,

year number 4.

month NUMBER (2),

parent_id NUMBER (12)

);

-Create/recreate primary, unique and foreign key constraints

change the table1 table:

Add primary key constraint PK_TABLE1 (ID);

change the table1 table:

Add the foreign key constraint FK_TABLE1_PARENT (PARENT_ID)

reference TABLE1 (ID);

data: