Complex transformations in SELECT

Question

Hi everyone,
I have a pretty complex situation where I have to apply transformations on data in a SQL SELECT Statement. The thing is that I have absolultely no idea how to handle this.
Let's imagine I have an hierarchy table, like an organizational hierarchy.
CREATE TABLE organisations
   organisation_id number(10) PRIMARY KEY,
   parent_organisation_id number(10),
   name_en varchar2(255),
   name_nl varchar2(255),
   CONSTRAINT fk_org_to_par_org
      FOREIGN KEY (parent_organisation_id)
       REFERENCES organisations(organisation_id)
INSERT INTO organisations VALUES (1, null, 'Top organization 01', 'Top organizatie 01');
INSERT INTO organisations VALUES (11,   1, 'Organization 011', 'Organizatie 011');
INSERT INTO organisations VALUES (12,   1, 'Organization 012', 'Organizatie 012');
INSERT INTO organisations VALUES (2, null, 'Top organization 02', 'Top organizatie 02');
INSERT INTO organisations VALUES (21,   2, 'Organization 021', 'Organizatie 021');
INSERT INTO organisations VALUES (22,   2, 'Organization 022', 'Organizatie 022');
INSERT INTO organisations VALUES (3, null, 'Top organization 03', 'Top organizatie 03');
INSERT INTO organisations VALUES (31,   3, 'Organization 031', 'Organizatie 031');
INSERT INTO organisations VALUES (32,   3, 'Organization 032', 'Organizatie 032');
INSERT INTO organisations VALUES (321, 32, 'Organization 321', 'Organization 321');
INSERT INTO organisations VALUES (322, 32, 'Organization 322', 'Organization 322');And I have to apply some transformations in the SELECT. These transformations are store in a table. I have a few different transformations. I just put 3 here as examples:
CREATE TABLE transformations
  transform_id number(10),
  transform_type CHAR(1),
  organ_id  number(10),
  organ_name varchar2(255),
  parent_organ_id number(10),
  replace_name varchar2(255),
  with_children char(1)
INSERT INTO transformations(transform_id, transform_type, organ_name, parent_organ_id, organ_id)
     VALUES (1, 'C', 'Created NODE', 1, -9);
INSERT INTO transformations(transform_id, transform_type, organ_name, replace_name)
     VALUES (2, 'R', 'Organization 021', 'Organization 021 renamed');
INSERT INTO transformations(transform_id, transform_type, organ_name, replace_name, with_children)
     VALUES (3, 'R', 'Organization 032', 'Organization 032 Renamed', '1');The first transformation is to create a node under an existing one. The second transformation example is to rename a node. The pattern is actually a regular expression and it will use regexp_replace (to be able to add in the begining, change, add in the end...). The 3rd example is a rename too but all the children of a node should also be renamed with the same pattern.
I need a result like this one: (I indented to read more easilly)
1, ,Top Organisation 01, Top organizatie 01    
  -9, 1, Created Node, Created Node     
  11, 1, Organization 011, Organizatie 011
2, ,Top Organisation 02, Top organizatie 02    
  21, 2, Organization 021 renamed, Organization 021 renamed     
  22, 2, Organization 022, Organizatie 022
3, ,Top Organization 03, Top organizatie 03    
  31, 3, Organization 031, Organization 031    
  32, 3, Organization 032 Renamed, Organization 032 Renamed
    321, 32, Organization 032 Renamed, Organization 032 Renamed
    322, 32, Organization 032 Renamed, Organization 032 Renamed I was thinking if it was possible to apply these modifications on a SELECT clause without creating another table and refreshing it by parsing the initial table? Any suggestions?
Thanks,

Answer 1

Hi Frank, 
I have been thinking for days on this problem... That's probably why I can't explain it correctly. SO let me try again. Let's restart evrything and recreate the ORGANISATIONS table.
set define off;
DROP TABLE ORGANISATIONS;
DROP TABLE TRANSFORMATIONS;
CREATE TABLE organisations 
   organ_id number(10) PRIMARY KEY, 
   parent_organ_id NUMBER(10), 
   organ_name varchar2(255), 
   CONSTRAINT fk_org_to_par_org 
      FOREIGN KEY (parent_organ_id) 
       REFERENCES organisations(organ_id)
INSERT INTO organisations(organ_id, parent_organ_id, organ_name) 
     VALUES (11, NULL, 'Sales');
INSERT INTO organisations(organ_id, parent_organ_id, organ_name) 
     VALUES (23, NULL, 'R&D');
INSERT INTO organisations(organ_id, parent_organ_id, organ_name) 
     VALUES (71, NULL, 'Logistics');
INSERT INTO organisations(organ_id, parent_organ_id, organ_name) 
     VALUES (87, 71, 'Transport');Once you executed the previous statements, you have the following result:
SELECT organ_id, parent_organ_id, CAST(LPAD(' ', LEVEL * 2) || organ_name as VARCHAR2(25)) AS a
  FROM organisations
 START WITH parent_organ_id IS NULL
CONNECT BY PRIOR organ_id = parent_organ_id;
ORGAN_ID               PARENT_ORGAN_ID        A                         
11                                              Sales                   
23                                              R&D                     
71                                              Logistics               
87                     71                         Transport             The organisation has 3 root department: Sales, R&D and logistics. The logistics department has a child department called transport.
Now, the previous tree comes from a remote database but that's not the point. It can not be used as provided. I have to apply some modification to that tree before I can use it. That's why I creat the TRANSFORMATIONS table. 
CREATE TABLE transformations
  transform_id number(10), 
  transform_type CHAR(1), 
  organ_id  number(10), 
  organ_name varchar2(255), 
  parent_organ_id number(10),
  replace_name varchar2(255)
);Now, every record inserted in that table is a transformation to apply to the original tree. One of the transformation I need to apply is to create organisations. 
For example:
INSERT INTO transformations(transform_id, transform_type, organ_id, organ_name, 
                            parent_organ_id)
      VALUES(  35, 'C', -19, 'Local sales', 11);The previous transformation means: Create a new organisation named 'Local Sales' under the Sales organization. Let's review the different values in the INSERT statement. The transform_ID is a generated ID. We don't really care about it.The transform_type 'C' means create an organization (node). The -19 is the ID that that new organization should have. 'Local Sales' is the name of that new organization and 11 is his parent id. That's with this parent-id that I know where to put the new node in the tree.
After applying this transformtation, the result should be:
ORGAN_ID               PARENT_ORGAN_ID        A                         
11                                              Sales                   
-19                    11                         Local sales 
23                                              R&D                     
71                                              Logistics               
87                     71                         Transport             Of course, I can have other CREATE transformations:
INSERT INTO transformations(transform_id, transform_type, organ_id, organ_name, 
                            parent_organ_id)
      VALUES(  64, 'C', -23, 'Foreign sales', 11);
INSERT INTO transformations(transform_id, transform_type, organ_id, organ_name, 
                            parent_organ_id)
      VALUES(  66, 'C', -26, 'South America', -23);
INSERT INTO transformations(transform_id, transform_type, organ_id, organ_name, 
                            parent_organ_id)
      VALUES(  97, 'C', -13, 'IT', null);After applying these transformations, the result should be :
ORGAN_ID               PARENT_ORGAN_ID        A                         
11                                              Sales                   
-19                    11                         Local sales 
-23                    11                         Foreign sales 
-26                    -23                          South America 
23                                              R&D                     
71                                              Logistics               
87                     71                         Transport             
-13                                             ITI hope this is clear now? I wanted to know if it's possible to apply such transformations directly in the SELECT statement without having to code a PL/SQL package that does it for me?
Now the problem is that I can have other type of transformations. For example, some nodes should be renamed. 
For example, if I insert this transformation:
INSERT INTO transformations(transform_id, transform_type, organ_id, replace_name)
      VALUES(  60, 'R', 11, 'Sales department');Instead of returning the name 'Sales' for the organization with ID = 11, it should return 'Sales department'. The 'R' transformation type means rename and the org_id is the organization to rename and the new name is supplied using replace_name.
The most difficult transformation is probably the next one. It defines nodes to retrieve by organization. For example, let's add the following data:
INSERT INTO organisations(organ_id, parent_organ_id, organ_name) 
     VALUES (50, NULL, 'Security');
INSERT INTO organisations(organ_id, parent_organ_id, organ_name) 
     VALUES (51, 50, 'Building');
INSERT INTO organisations(organ_id, parent_organ_id, organ_name) 
     VALUES (80, 51, 'Parkings');
INSERT INTO organisations(organ_id, parent_organ_id, organ_name) 
     VALUES (81, 80, 'Parking 1');
INSERT INTO organisations(organ_id, parent_organ_id, organ_name) 
     VALUES (82, 81, 'Parking 1 - Level 0');
INSERT INTO organisations(organ_id, parent_organ_id, organ_name) 
     VALUES (52, 50, 'Software');
INSERT INTO organisations(organ_id, parent_organ_id, organ_name) 
     VALUES (52, 50, 'Other kind of security');AS you can see in the following result, the Security organization has 4 levels below.
SELECT organ_id, parent_organ_id, LEVEL, CAST(LPAD(' ', LEVEL * 2) || organ_name as VARCHAR2(30)) AS a
  FROM organisations
 START WITH parent_organ_id IS NULL
CONNECT BY PRIOR organ_id = parent_organ_id;
ORGAN_ID               PARENT_ORGAN_ID        LEVEL                  A                              
11                                            1                        Sales                        
23                                            1                        R&D                          
50                                            1                        Security                     
51                     50                     2                          Building                   
80                     51                     3                            Parkings                 
81                     80                     4                              Parking 1              
82                     81                     5                                Parking 1 - Level 0  
71                                            1                        Logistics                    
87                     71                     2                          Transport                  
 9 rows selected 
 I also should be able to define with a transformation the number of level to retrieve for a specific node. 
ALTER TABLE transformations 
  ADD max_children number(10) default 4;
INSERT INTO transformations(transform_id, transform_type, organ_id, max_children)
      VALUES(  20, 'L', 50, 2);In the previous transformation, I specify a 2 level for the organization 50. In this case, the final result should only return 2 level under the organization 50. So instead of having 5 levels as in the previous result, I should have
ORGAN_ID               PARENT_ORGAN_ID        LEVEL                  A                              
11                                            1                        Sales                        
23                                            1                        R&D                          
50                                            1                        Security                     
51                     50                     2                          Building                   
80                     51                     3                            Parkings                 
71                                            1                        Logistics                    
87                     71                     2                          Transport                  As you can see, there are a lot of different type of transformations that should be done and I don't think SQL is intented for such things?
Thank you again for your help Frank, I really appreciate