TOS - Help with Data transformation logic

[Anonymous]

Hello everyone, I have to make an etl and I stuck on a logic of transformation and maybe someone can help me. Simplifying a lot now my file is structured as below:

ACTIVITY	IN	OUT
A	1	2
A	2	3
A	3	4
A	6	7

The logic is that if there are 2 or more rows that are sequential ( OUT corresponding with IN ), I have to aggregate, keeping the minimum IN and the maximum OUT. So the results has to be:

ACTIVITY	IN	OUT
A	1	4
A	6	7

If there were just the first 3 rows, it would be easy using a tAggregate, but I don't know how to tell Talend that has to aggregate only sequential ones. (IN and OUT in reality are dates) 

[Anonymous]

Basically, you should think of creating partition buckets.  Once you have your partition buckets, you just do a min on IN and a max on OUT on the bucket group and you will have the answer.

Here is an example using integers.  But you can use same logic for dates. I have done it with a tJavaFlex and some code as it is the fastest if your data is sorted on IN and is just like below.