In dev: Designer Patterns for Pentaho Data Integration
Name Schema in Steps:
In Input tables never add a name schema:
SELECT id, name FROM MYDATABASE.MYTABLE , this is no good!
, the name schema as add in connection , ans you command SQL Correct is :
SELECT id, name FROM MYTABLE, this is good!
WHY ?
In case of database migration you will only change one location and all your changes or Jobs continue to work.
Share Connections
Just keep a shared connection while you are working on a project with that connection, and on each etl just keep the valid icons for the particular etl that is using them.
Why ?
If you leave everything shared and you have to send your etl to a client, all the links are in the last edited etls.
Look this , this ETL use only dw_responsiva :
Command SQL in Steps:
Write commands SQL em UPPERCASE!
Input ou Output file in Operation Systems
Never add a absolute path in steps, add a relative path or create a VARIABLE ( kettle.properties or other forms)
Look a sample the correct form :
Why ?
If the environment shortcut change will have to maintain its transformation or job
Name Of Transformations and Name Of File
The property name must be equal to the filename KTR file.
Sample
Property Name Of Transformation : LoadClients
Name Of Filename KTR : LoadClients.ktr
