HI, is it possible to copy a row from the Input0_ProcessInputRow into another variable of a synchronous script component. I have a script component that has three outputs. What I would like to do is the following:
At first iteration, copy entire row into a variable.
For the other rows, is business key cols is different from the previous row (the one copied at previous iteration) or type 2 columns :
1-direct previous row (from the variable) to second.output
2- direct current row to the first output
3- copy the current row into the variable
If the business key cols are the same and type 2 cols are the same:
1-Direct row to the third output
My problem is that when I declare a variable as Input0Buffer and then copy the current row in it, at each subsequent iterations, the variable is updated. What I want to do is to keep the value of the previous row. I know I can do it b are differenty copying the content of all columns into separate variables but since some script components will have 150+ columns, I would like to find a simpler way to do it (like a copy of the structure of the input buffer).
The reason I use a script component is that the lookups are not updated dynamically, the rows I need to process may not exist in the target table and once inserted, they are not retreived subsequently by the buffer.
Thank you,
Cco
ccote wrote:
My problem is that when I declare a variable as Input0Buffer and then copy the current row in it, at each subsequent iterations, the variable is updated. What I want to do is to keep the value of the previous row. I know I can do it b are differenty copying the content of all columns into separate variables but since some script components will have 150+ columns, I would like to find a simpler way to do it (like a copy of the structure of the input buffer).
You should not use the buffer that was passed to you outside of the ProcessInputRow call - the data flow engine manages the buffer lifetime, and by trying to use it outside of this call you are breaking the rules.
Try copying the columns to another data structure, e.g. an array.
|||HI, thank you for your reply. I knew I could copy the row's column to another structure (array, list, variables) but I was wondering If I could copy the whole row into a kind of "row" structure and be able to later direct it to an output. It would have been great since as I said, some of my tables have 150+ columns. I guess, the only way I can manage this is to actually add type1 and type 2 columns to my output and accumulate previous row type1 and type2 attributes into variables. So when the business key changes, I would copy variables into current row and have a conditionnal split that further filter out if any action have to be taken.
I could also use asynchrounous script but there I need to recreate all my 150+ table structure into output buffers. Actually, the real problem is that the lookup transform cache is not updated as new rows are added to the target table. That is why I need to do some custom code like this.
Thank's again,
Ccote
Phil Brammer wrote: