Remove duplicated columns Transformation Script

Hi everyone
i need a script to remove the duplicated column in the relational model
can any one help me on that?
Thanx in advance.

Try this
columnName = 'Column_1';
tables = model.getTableSet().toArray();
for (var t = 0; t<tables.length;t++){
table = tables[t];
columns = table.getElements();
size = table.getElementsCollection().size();
for (var i = 0; i < size; i++) {
    column = columns[i];
    if (column.getName() == columnName){
        column.remove();
table.setDirty(true);
Replace 'Column_1' with your column name

Similar Messages

Sql Developer Data Modeler 3.0 EA1: Custom transformation script

What is the object model used by:
Tools -> Design Rules -> Transformations
From looking at the examples (and guessing) I have managed to create ID attributes on all my entities, but how do I set its sequence number or set it as the primary key?
entities = model.getEntitySet().toArray();
for (var e = 0; e<entities.length; e++)
entity = entities[e];
att = entity.createAttribute();
att.setName('ID');
- Marc de Oliveira

content of XML files:
1) class described, parent class if there is one
2)<property .../>
describes property - set and get method, data type of property, default value
3) <collection ... />
describes collection of objects belonging to class - 5 collections are defined in Table.xml - columns, indexes, table check constraints, column groups and spatial definitions
- get method - to get collection
- create item method - method of described class used to create item for that collection; so for columns collection table.createColumn() will create column in table
- other methods - add, move, remove if are defined
4)other meta data
definitions in 1) .. 4) are used by Oracle SQL Developer Data Modeler, so it'll be wrong to modify those xml files
5)<roproperty ../>
Data Modeler doesn't parse/process these definitions, we decided to not filter them out however you should not rely that much on them
6)some junk left
For each object you can set your own properties and they will be persisted (also included in compare/merge in versioning) together with other information for object - look at "template table" example in transformation scripts. Here are methods you can use:
void setProperty(String key, String value);
     String getProperty(String key);
     boolean hasProperty(String key);
     boolean hasProperty(String key, String value);
     void removeProperty(String key);
     void clearProperties();
     Iterator getPropertyNames();
Philip

Transformation script and methods

Hi !
I'm writing transformation scripts and I want to know where I can find a complete list of method ?
In API documentation in \xmlmetadata folder, I can find getter and setter for each class, but some other methods seem to exist. Examples:
on Column class, a copy() method, a remove() method
on FK , an addToAllDPVs() method (see How to use transformation script to add FK to relational model
model.getAppView.log()
Typically, does a copy() method exist for Table class ?
Where can I find it ?
Thanks
Regards

You just have to define a method in the private section of the header of the coding. Just scroll up to the top when creating a routine.
Please see also the general help on private methods.
<a href="http://help.sap.com/saphelp_nw2004s/helpdata/en/c3/225b6554f411d194a60000e8353423/content.htm">http://help.sap.com/saphelp_nw2004s/helpdata/en/c3/225b6554f411d194a60000e8353423/content.htm</a>
Cheers
SAP NetWeaver BI Organisation

Advice for custom transformation script to change table FK Associations

Hello,
I am attempting to write a transformation script that will cycle through tables in a large model and modify some
FK associations to change the remote table in the FK. As an aside, the reason for this is that our logical model
has logical entities that represent many small lookup code domains. However our relational/physical model has a
single utility code repository physical table which contains all these code domains. So part of the transformation
from our logical model to relational model is to swing all the FK associations for these lookup code domains
into the common code table. Then the tables representing the lookup domains can be hard deleted.
In any case, I'm looking at the
oracle.dbtools.crest.model.design.relational.FKIndexAssociation
or the
oracle.dbtools.crest.model.design.relational.Index
to see if one of these can be manipulated to perform the transformation I want to accomplish.
When I perform the equivalent process manually via property sheets there are several ways of accomplishing this.
a) The neatest method seems to be to modify the existing FK in place by clicking on the FK in the relational diagram and then changing
the property called 'PK / UK Index'. Change that to the PK of the utility code repository table that is replacing the original
table (representing the lookup domain). A side effect of this results in the mapping of the FK
columns being nulled out. So they need to be remapped to the original columns (which are retained).
So in two steps the existing FK is detached from the original table, reattached to the new table and then
the FK columns are reused in place for the new FK mapping.
b) The brute force method is to delete the original FK and then create an entirely new FK to the new table. There are
multiple dialogs which allow you to retain the original FK columns and reuse them in the new FK.
I realize this is a somewhat complex transformation but I'd appreciate some algorithmic advice on which path to take
to try to accomplish this. In particular, is there a scripting equivalent to the first process above?
I don't even know if the FK that is represented on the relational diagram is represented by the FKIndexAssociation or Index
class in the scripting object model. In other words, as I loop through the tables in the relational model what would be my
starting point:
fks = table.getFKAssociationsList();
or
fkIndexes = table.getAllInds_FKeyInds();
or something else. The names of the properties in the user interface don't always match with the scripting properties so
it's a little tricky to figure this stuff out.
Once I get access to the object that is equivalent to the FK in the diagram/property sheet, what
is the best way to manipulate it via scripting. How can I change the parent/remote end of the FK to point to the new table.
Or if I have to delete the original FK and then recreate it - is there a way to save the original FK columns and reuse them
while creating the new FK?
One other question. What is the best way to delete tables from a relational model. I note that there is an undocumented (?) remove()
method that is available on Table. It is probably on a superclass so other things can be removed as well.
When I try to use this method in a loop the script throws up dialogs similar to the dialogs that appear while performing a delete
manually via the UI. Is there any way to intercept and respond to the dialogs programmatically?
So for instance if I get a dialog popup like
Do you want to delete generated FK columns 'FOO' in table 'BAR'?
is there a way to intercept that dialog in the script and then answer Yes or No to it via the script?
If it turns out to be too difficult to perform this type of transformation via scripting the fallback
is to perform the transformation on the generated DDL file. But it would be cleaner to be able
to perform the transformation in the DM Relational model if possible.
Any advice appreciated.
Rgds, BP

Philip,
This is good info. I was able to get a rough version of my transformation script to work.
However I'd like to clean it up a bit.
In your notes it is not clear how the FKIndexAssociation and associated Index are related.
Is there a getter on FKIndexAssociation that references the associated Index?
It looks like the method fk.changeKeyObject() takes an Index object as its first parameter.
What would be the cleanest way to extract the PK Index from the Code table that I want
to point to? I used something like:
cdTable = model.getTableSet().getByName("CD_REPOSITORY");
cdPkIndex = cdTable.getIndexByName("CD_REPOSITORY_PK");
I used the latter method because it's the only one that returns an object of type Index
which is required as input to FKIndexAssociation.changeKeyObject(). However this method
requires the name of the index. That's fine but I'd like to just grab the PK's supporting index
directly if possible (without knowing the name).
I'm curious if there is a getter on Table or FKIndexAssociation that can just grab the index of the PK of
a table directly?
I don't see such a getter on FKIndexAssocation. On the Table class there are many getters that might be
of some use but the only one I could get to work was getIndexByName() as shown above.
I tried to use getIndexes() and then use isPk() on the result of that in a loop but that didn't return any result.
It looks like getters getIndexes(), getPKs(), getUKeys() return collections of DesignObject which are just the
superclass of FKIndexAssociation which is not what I need (i.e. which IS the actual underlying Index object).
So I guess the general question is what is the best way to get access to the Index objects in two scenarios.
First for the PK and UKs (typically on the 'parent-referenced' table) and secondly for FKs (typically on the 'child-referencing' table).
If I were imagining how this would work I would imagine getters on the Table for the former and a getter
on the FKIndexAssociation for the latter. But that's not how it works.
Also, as an aside - this 'Index' is somewhat confusing since these seem to be 'potential indexes' and not true database indexes.
These 'indexes' seem to be more closely related to database constraints and not actual indexes. I know that Primary Keys are
implemented in Oracle via a unique index but I'd like to confirm that the class
oracle.dbtools.crest.model.design.relational.Index
is not the same thing as a DDL database Index.
Secondary question. I notice when I run the script which loops through all the tables and redirects the FKs - the
diagram does not refresh. I had to save the model and then reload it for the graphical representation
of the FKs to change and redirect to the new table.
Is there a way to force the refresh to occur right away?
I've noticed that other transformation scripts I've written HAVE resulted in the diagram being refreshed
right away (even while the script is running). In this case it does not.
Rgds, BP

Easiest way to write transformation scripts

Hey,
I need to write quite a lot of transformation scripts. As I'm not a expert in javascript and java, I was wondering what's the easiest and quickest way to start creating those scripts. I found it quite annoying in Oracle SQL Modeler to switch between the library and the execution of the scripts. For some scripts I need interaction with the database and for other ones I need to ask the user for some parameter values.
I've already tried the following with no real success:
- Creating a project in JDeveloper
- In Rhino command line, but I'm missing the Oracle SQL Modeler libraries
- Opening 2 times Oracle SQL Modeler with the libray editing window and the transformation running window, the transformation running window does not see the changes in the other one.
I've searched the internet, but I did not found any usefulle information about this subject.
Any help would be appreciated.
Thanks.
Dominique

Hi Dominique,
I write my scripts in "Custom transformations scripts" dialog - "Tools>Design rules>Transformations".
Opening 2 times Oracle SQL Modeler with the libray editing window and the transformation running window, the transformation running window does not see the changes in the other one.
In "Custom transformations scripts" dialog you can write helper functions and execution code that do something and utilize those helper functions. Well if you want to reuse those helper functions in another
transformation you need to copy them in new transformation. Libraries are here to help to overcome that limitations. Normally you write and test your helper functions in "Custom transformations scripts" dialog
and then put them in library. Also you put there a function that corresponds to the code you have in single transformation script and that can be invoked as single transformation. Data Modeler doesn't parse the library, you need to define
functions that can be executed directly - that in "Functions/Methods" list of "Custom libraries" dialog. Functions exposed in that list are available for selection in "Custom transformations scripts" dialog. So what you expose is what you can use.
Quick example - DM comes with script called "Tables to lower case - Rhino "
tables = model.getTableSet().toArray();
for (var t = 0; t<tables.length;t++){
table = tables[t];
name = table.getName().toLowerCase();
table.setName(name);
columns = table.getElements();
size = table.getElementsCollection().size();
for (var i = 0; i < size; i++) {
    column = columns[i];
    cname = column.getName().toLowerCase();
    column.setName(cname);
table.setDirty(true);
keys = table.getKeys();
for (var i = 0; i < keys.length; i++) {
key = keys[i];
    if(!key.isFK()){
     kname = key.getName().toLowerCase();
     key.setName(kname);
   }else{
     kname = key.getFKAssociation().getName().toLowerCase();
     key.getFKAssociation().setName(kname);
     key.getFKAssociation().setDirty(true);
here is the same script put in library:
function lower_case(){
tables = model.getTableSet().toArray();
for (var t = 0; t<tables.length;t++){
table = tables[t];
name = table.getName().toLowerCase();
table.setName(name);
columns = table.getElements();
size = table.getElementsCollection().size();
for (var i = 0; i < size; i++) {
    column = columns[i];
    cname = column.getName().toLowerCase();
    column.setName(cname);
table.setDirty(true);
keys = table.getKeys();
for (var i = 0; i < keys.length; i++) {
key = keys[i];
    if(!key.isFK()){
     kname = key.getName().toLowerCase();
     key.setName(kname);
   }else{
     kname = key.getFKAssociation().getName().toLowerCase();
     key.getFKAssociation().setName(kname);
     key.getFKAssociation().setDirty(true);
And lower_case is exposed as function in "Function/methods" list
Philip

Programmatically creating Transformation Script Component

Does anyone have any examples of programmatically creating a Transformation Script Component (or Source/Destination) in the dataflow? I have been able to create other Transforms for the dataflow like Derived Column, Sort, etc. but for some reason the Script Component doesn't seem to work the same way.
I have done it as below trying many ways to get the componentClassId including the AssemblyQualifiedname & the GUID as well. No matter, what I do, when it hits the ProvideComponentProperties, it get Exception from HRESULT: 0xC0048021
IDTSComponentMetaData90 scriptPropType = dataFlow.ComponentMetaDataCollection.New();
scriptPropType.Name = "Transform Property Type";
scriptPropType.ComponentClassID = "DTSTransform.ScriptComponent";
// have also tried scriptPropType.ComponentClassID =typeof(Microsoft.SqlServer.Dts.Pipeline.ScriptComponent).AssemblyQualifiedName;
scriptPropType.Description = "Transform Property Type";
CManagedComponentWrapper instance2 = scriptPropType.Instantiate();
instance2.ProvideComponentProperties();
Any help or examples would be greatly appreciated! Thanks!

I am working on the creating script component programmatically.
The code has been created is shown in the following,
Dim SourceCodeValue(3) As String
Dim scriptPropType As PipeLineWrapper.IDTSComponentMetaData90 = _
CType(CType(dataflowTask, TaskHost).InnerObject, PipeLineWrapper.MainPipe).ComponentMetaDataCollection.New()
scriptPropType.Name = "Transform Property Type"
scriptPropType.ComponentClassID = "{BF01D463-7089-41EE-8F05-0A6DC17CE633}"
scriptPropType.Description = "Transform Property Type"
Dim instance2 As PipeLineWrapper.CManagedComponentWrapper = scriptPropType.Instantiate()
instance2.ProvideComponentProperties()
SourceCodeValue(0) = "dts://Scripts/" & "ScriptComponent_9ddf9eba6f5b488f9710b95e4d7e7c74" & _
"/" & "ScriptComponent_9ddf9eba6f5b488f9710b95e4d7e7c74" & ".vsaproj"
SourceCodeValue(1) = My.Resources.ScriptComponentProjFile
SourceCodeValue(2) = "dts://Scripts/" & "ScriptComponent_9ddf9eba6f5b488f9710b95e4d7e7c74" & _
"/ScriptMain.vsaitem"
SourceCodeValue(3) = My.Resources.ScriptComponentCode
Dim anIDTSDesigntimeComponent90 As Microsoft.SqlServer.Dts.Pipeline.Wrapper.IDTSDesigntimeComponent90 = _
CType(instance2, Microsoft.SqlServer.Dts.Pipeline.Wrapper.IDTSDesigntimeComponent90)
anIDTSDesigntimeComponent90.SetComponentProperty("SourceCode", SourceCodeValue)
anIDTSDesigntimeComponent90.SetComponentProperty("PreCompile", False)
When debugger reach to “anIDTSDesigntimeComponent90.SetComponentProperty("SourceCode", scriptValue)”, NullReferenceException. “Use the “New” keyword to create an object instance” was shown.
Could you please give me some help or suggestion?
Thank you so much in the advance

Subview formatting using transformation script

Hi,
We have 200+ subviews with tables consisting of roughly 100 columns a piece. After the models were created, we decided that we would only show the PK and FK columns and hide the rest. I was able to write a transformation script to only show those columns, but now my tables are showing up as if they have 100 columns, but only show 2. What do i need to do in the transformation script to "Best Fit" the size of each of the tables, and then set auto layout - layout 1 for the diagrams.
Thanks
diag = model.getAllDiagrams();
for(it = diag.iterator(); it.hasNext(); ){
      cur = it.next();
      log(cur.getName());
      cur.setShowAllDetails( false);
      cur.setShowKeys( true);
      cur.setShowLegend(false);
     //TODO
     //Some loop to "BEST FIT" all components in the diagram
     //TODO
     //run auto layout to arrange the diagram
     //maybe rearrangeNewDiagram() or setLayout(LayoutManager mgr)?
      cur.setDirty( true);
def log(String str){
oracle.dbtools.crest.swingui.ApplicationView.log(str);

Perfect, that's what I needed. Thanks
diag = model.getAllDiagrams();
for(it = diag.iterator(); it.hasNext(); ){
        cur = it.next();
        log(cur.getName());
        cur.getPlaceHolder().setVisible(true);
        cur.setShowAllDetails( false);
        cur.setShowKeys( true);
        cur.setShowLegend(false);
        cur.resizeTables();
        cur.resizeViews();
        cur.rearrangeDiagram2(1, true, 'Layout 1');
        cur.setDirty( true);
def log(String str){
    oracle.dbtools.crest.swingui.ApplicationView.log(str);

Removing Duplicated PLd

Hi All,
How can I remove duplicated PLD's? I tried to duplicate a PLD from Invoice to purchase order but now i want to remove that duplicated PLD. How can that be done?
Thanks,
Joseph

Hi Istvan,
Thanks.
Joseph

How to remove all columns and cells in numbers

how to remove all columns and cells in numbers

Click on the Table's icon in the Sheets list. Press delete.
Done.
Regards,
Barry

HT202905 applescript to remove certain columns in numbers version 3.5.x?

How do I tell applescript to remove certain columns in numbers version 3.5.x?

Via something like:
tell application "Numbers"
delete column 2 of table 1 of sheet 1 of document 1
end tell
To effectively AppleScript Numbers.app you need to understand its hierarchy - that is, documents contain a number of worksheets. Each worksheet has a number of tables (could be 1, could be many) and therefore you need to provide a complete reference to identify the column you want to delete.

Removing a column

It looks as if there is no way to hide a column in Discoverer Plus. Please confirm this. Secondly, if the only way to hide a column is to remove the column, if this field is used in a condition, that condition is automatically removed from the worksheet. Is there any way to set up a warning that the condition has been unchecked?

It's under the Sort option or tab. Add the column you want to remove to the sort and then I think you click or choose the 'Hide' or 'Hidden' option next to the column.

Error TypeError: null has no such function "getAppView" logical transform script version 4.1.0.866

Hello,
I am using Oracle Data Modeler version 4.1.0.866 and when I created a transform script in object logical, motor Oracle Nasnhorm give the error TypeError: null has no such function "getAppView".
The code is:
var guipkgs = JavaImporter(java.lang, java.awt, java.awt.event, java.awt.Window, Packages.javax.swing ,java.lang, Packages.javax.swing.border);
with (guipkgs) {
    log_rep = new Packages.oracle.dbtools.crest.swingui.LogReport(model.getAppView());
    log_rep.setName("CONTADOR");
    var cont = 0;
    var text = "";
entities = model.getDesign().getLogicalDesign().getEntitySet().toArray();
for (var t = 0; t<entities.length;t++){
entity = entities[t];
    Name = entity.getName();
    cont = cont + 1;
    text = text + " " + Name + "\n";
text = text + cont + "\n";
            log_rep.setLog(text);
    log_rep.showModalDialog(new java.awt.Dimension(1300, 550));
Already you can not use log_rep = new Packages.oracle.dbtools.crest.swingui.LogReport(model.getAppView());? Is there any alternative?
This error does not come with previous versions.
Thanks.

Hello,
I modified the script add following at the beginning of script:
model = model.getDesign().getLogicalDesign();   and carry on the error TypeError: null has no such function "getAppView".
var guipkgs = JavaImporter(java.lang, java.awt, java.awt.event, java.awt.Window, Packages.javax.swing ,java.lang, Packages.javax.swing.border);
with (guipkgs) {
    model = model.getDesign().getLogicalDesign();
    log_rep = new Packages.oracle.dbtools.crest.swingui.LogReport(model.getAppView());
    log_rep.setName("CONTADOR");
    var cont = 0;
    var text = "";
entities = model.getDesign().getLogicalDesign().getEntitySet().toArray();
for (var t = 0; t<entities.length;t++){
entity = entities[t];
    Name = entity.getName();
    cont = cont + 1;
    text = text + " " + Name + "\n";
text = text + cont + "\n";
            log_rep.setLog(text);
    log_rep.showModalDialog(new java.awt.Dimension(1300, 550))
Any idea?
Thanks

1.5PROD/EA3/EA2 - Issues removing indexed column through Edit Table dialog

The Edit Table dialog does not cope well with removing a column which is used in indexes.
If you simply remove the column and hit OK, you get a Validation Failed error saying that "Column <column> is not a column of table <table>". Given that you are trying to remove the column, this is confusing.
Most times, you can navigate to the indexes but the column no longer shows in the index which contained (and all other columns in the index appear to be removed as well). At least you can remove the index and then accept the change. However, some times (couldn't reproduce consistently), the index didn't display properly (no name displayed) and I couldn't leave the record without entering a dummy name before removing it.
Shouldn't the behaviour be equivalent to the "corresponding" alter table drop column command which automatically drops the indexes referring to the dropped column? I know that it isn't always nice to "silently" remove an index or indexes, but couldn't we display a dialog that indicates which indexes we are going to drop to warn the user?
theFurryOne

No response to this, but it is still a problem in EA3.
I assume it is not intended that you cannot drop an indexed column through the Edit Table dialog, without having to first drop the index - it is certainly not required in direct SQL and the Drop Column via the Table context menu works as well (as that simply generates the SQL).
theFurryOne

Derived column transformation from varchar(8) YYYYMMDD to DATETIME column

It seems I have tried everything, but all I need to do is change a column from varchar(8) to DATETIME. Some lines are blank, zero length strings, and some have dates such as 20140829. All the rows go to an exception table or it just bombs altogether. The
only thing I can get to run is NULL(DT_DATE), but obviously, that won't work in the end.
I tried LTRIM(RTRIM(MyDateColumn) == "" ? (DT_DATE)"1900-01-01" : (DT_DATE)(SUBSTRING(MyDateColumn,1,4) + "-" + SUBSTRING(MyDateColumn,5,2) + "-" + SUBSTRING(MyDateColumn,7,2))
I got truncation errors on that.
Sometimes I got an error: An error occurred while attempting to perform a type cast.
I tried DT_DATE, DT_DBDATE and DB_DBTIMESTAMP
I even tried putting GETDATE() in there or a literal date such as "2014-08-29" like (DT_DBDATE)"2014-08-29"
This wouldn't even compile: MyDatCeolumn == "" ? NULL(DT_DBDATE) : (DT_STR,10,1252)((DT_DBDATE)MyDateColumn)
I tried many variations on these. I either got truncation errors or the transformation would turn red and stop, or the expression wouldn't work (turns red when you leave the row).
Review:
Source MyDateColumn VARCHAR(8) NULL; Target MyDateColumn DATETIME; Source values in format of "" (blank) or "20140829". The blanks should end up null, or maybe a literal date like Jan 1, 1900, and the actual dates should be transformed
to DATETIME.
Should not be that hard. Thanks in advance for help.

Hi duanewilson,
Please create a Derived Column transformation and a new column, then use the following expression for the expression field:
(DT_DATE)(SUBSTRING([MyDateColumn],1,4) + "-" + SUBSTRING([MyDateColumn],5,2) + "-" + SUBSTRING([MyDateColumn],7,2))
If you have any feedback on our support, please click
here.
Elvis Long
TechNet Community Support

Derived Column Transformation Editor and a Date

My vendor file used to have "N/A" in the Date field and I used a Derived Column Transformation Editor to feel this out with the following expression...
PAPER_STOP_DATE == "N/A" ? NULL(DT_WSTR,50) : PAPER_STOP_DATE
Now the vendor has also decided to pass nothing in there. So I honestly don't know if it's a "" or a <NULL>. It's coming from an Excel spreadsheet.
Is there any easy way to add "" to this expression and how? Or is there a way to determine if this field is a proper date format?
This is the date that exists in the cell...
03-aug-2014
After I determine if it is a "N/A", I am doing a Data Conversion on it and it is properly loading to our SQL Server Database.
Thanks in advance for your review and am hopeful for a reply.
ITBobbyP85

If i understand correctly you want to derived a column to NULL if the input "paper stop date" is "N/A" or "" or a NULL else the actual "paper stop date". Right ?
PAPER_STOP_DATE == "N/A" || TRIM(PAPER_STOP_DATE) == "" || ISNULL(PAPER_STOP_DATE) ? NULL(DT_WSTR,50) : PAPER_STOP_DATE
Regards, RSingh

Remove duplicated columns Transformation Script

Similar Messages

Maybe you are looking for