Creating Groups

[lrhatigan]

Hi, 

Any help in grouping a large data set using the requirement detailed in the example below would be greatly appreciated. 

The first four rows fall into the same group due to the fact that a record within the row (marked in bold) has been categorised previously. Row 5 contains a completely unique data set, therefore a new group is created, as has Row 7.

Essentially, if any value is repeated throughout Col A, B or C, which has already been grouped, it should inherit the same group.

RecordID	Col A	Col B	Col C	Combined	Outcome
1	a	1	y	a 1 y	Group 1
2	a	2	x	a 2 x	Group 1
3	b	3	z	b 3 z	Group 1
4	b	1	z	b 1 z	Group 1
5	d	20	t	d 20 t	Group 2
6	d	30	u	d 30 u	Group 2
7	e	50	p	e 50 p	Group 3
8	f	3	l	f 3 l	Group 1

[PhilipMannering · Accepted solution]

Hi Irhatigan,

If I understand correctly, I think record 3 might be wrong in your dummy data as all three fields are unique but you've still allocated it as group 1...

Think I cracked it eventually though (this was very challenging!). Take a look,

workflow

Let me know if this is what you're after.

Thx,

P

[lrhatigan]

Hi Philip, 

The solution provided has worked, thank you. I had to modify it slightly as my data includes null values, but this has provided a much faster way of grouping unique scenarios in a large data-set.

Record 3 contains 'b' in ColA which is also contained in Record 4 so therefore not unique.

Thanks again!