Cell content deduplication

KNIME Analytics Platform
1

 

How can I deduplicate content of two or more cells of two columns after or before aggregation?

example table:

Column A                                        Column B                                  Column C (aggregation)

Germany, München, Napoli      Rome, München, Milano        Germany, Rome, München, Napoli, Milano

Greece, Austria, Rhodes       Greece, Kavala, Austria, Vienna       Greece, Austria, Rhodes, Kavala, Vienna  

and so on...

 

P.s: I forgot to add that I can sole this problem by combining column aggregator and cell splitter nodes, but it is time-consuming and feels like using a cannon to kill a fly.

2

I just played around a few minutes. I can't think of any solution other than using at least three nodes:

  1. Column Aggregator (use Concatenate as aggregation method)
  2. Cell Splitter (use comma as delimiter)
  3. Column Aggregator (use Wildcard/Regex Selection of columns and Unique concatenate as aggregation method)

You can enable the option Remove aggregation columns of the Column Aggregator node to get rid of all the unnecessary columns created by the Cell Splitter node.

3

Hi ImNotGoodSry,

you are completely right. You need use those three to this job. On the other hand, I fell that there must be a easier  solution, maybe asn an option in one of the nodes (column aggregator might be the right place)

4

After column aggregator node, you can use the Ungroup node or split collection cell node.

simon.

5

Ungroup would put the values to the rows, I do not think that was the intent. Probably Split Collection Column?

6

Hi Aborg,

yes, split collection cell will put them into columns, whilst Ungroup will put them into rows. So boraster has two options of how he prefers the output.

i forgot to make clear to boraster too that in the column aggregator node, you need to aggregate by List, not concatenate, for you to be able to use these two nodes.

simon.

7

Both paths lead to same destination, but I wish there was a one-click solution for this simple problem :)

1 Like