When building my transform i sometimes need to go back and insert a new upstream transform items (e.g. a new col or, copy col) into my build. But when i do this it resets some of the downstream transforms.
As an example, if i have a rename transform at the end of my build , if i need to go back and insert an uptream transform e.g. Copy col , i lose all the mappings on the rename transform and then need to redo it, its like its been reset as I have disturbed the original build.
Whats the best way to avoid/limit this ? Is there an efficient way to insert a new transform upstream ?
ahhhh i think part of this must be how was adding these transforms, i was deleting the connection line instead of doing a right click and adding them…doh…
You are correct deleting the connection can lose information. You can mark the arrow and add then a new transformation.
You need to keep in mind that EDT internal the columns are addressed just on the columns number and that added column or change of columns order in the input will impact the logic. Therefore I recommend to use “Stack” directly after reading data and a second file with just the field names to ensure same order everytime. Read the docu or help on Stack.
When you delete a connection is nukes all the column related settings downstream. To avoid this you should try to insert and delete transforms without deleting connections:
If the column order in your input changes (e.g. an extra column is added at the start), this can mess things up. To avoid this you can Stack again the expected column order: