Full outer join

CloverDX Platform
1

Is there a way to do a full outer join, without joining the files twice? Am I missing something – I only saw left outer.

2

While we are in this topic, is it possible to have several out ports, 3 in particular: (if A and B are two files)
outport0: results of inner join between A and B
outport1: records in A but not in B (outer join on B)
outport2: records in B but not A (outer join on A)

… assuming join on some key.

Above is “nirvana” for data warehousing situations, reducing considerable processing.

Thanks
Akhil

3

That’ll be awesome!

In Data Warehousing situations where one routinely compares a data set (say employee) from one period to another to get inserts (new employees)/updates (changed employees) and deletes (deleted employees) , a full outer join can do all 3 in one shot. Very useful for doing type-1 and 2 versioning.

This way, one needs to join the 2 data sets only ones.

By the way, cloverEtl rocks! Great Job!!!

Do you have a target feature list/date for the next release?

Also, I am doing some bench marks comparing clover tasks with similar oracle/mySql tasks. I’ll publish it when I am done.

Best
Akhil

4

Hmm,

I have never seen such component in “big” ETL tools, but this is really easy one. I have finished the outer join option in MERGE_JOIN and this is just simple modification - based on matching driver&slave records send it to one of the three ports.

David.

5

Hi!

Your observation is correct. There is no full-outer-join as of version 1.1.4.
I will try to include it in version 1.1.5 as this is a simple addition. Only I have never thought of using it.

David.

6

AbInitio is one such tool, that’s where I was exposed to this component. This tool is a market leader in handling high volume ETL.