If you could please explain in plain text (possibly including a short sample set of data) what this stuff does, we should be able to help you.
Please do not expect everyone to be a SQL expert.
Can you please elaborate more on your requirement. Can you please clarify :
-What exactly you want to achieve
-Which trasformation you want to use.
Please share more details to understand on your requirement.
You needed all the fields from the table t and there is a derived column from the same table. Probably a two different SQ pipeline reading the same table t, one pipeline you can have Aggregator to do the count and then you might need to join the two pipelines using the columns col 2 & 3.
Achieving this with a single pipeline can also be done, I guess.
Relational Source Qualifier followed by a Sorter to ensure data are sorted according to partitioning keys.
An Aggregator to perform the COUNT(*).
A Joiner to perform the join (with Sorted Input) on the output of the Aggregator and the output of the Sorter.
If the original poster would please explain what this PARTITION BY does, we could give more details.