Bug: GroupBy Node Append Elements not working as expected

Feedback & Ideas
1

I have a list column and as aggregation I’m choosing “Append Elements” with “Filter missing collection elements” unchecked.

image

However the generated output list never contains any missing collection elements. Either I don’t understand the option correctly (Filter missing collection elements) or it isn’t working. In fact the option doesn’t have any effect on the output. This applies to KNIME 3.7.2 and KNIME 4 and possibly earlier versions.

See attached workflow for an example.

image
image1092×231 20.9 KB

Append Elements.knwf (8.2 KB)

2

The wording should be filter out missing…
If you realy would like to get only thw missing values use the row filter node…

3

I want all the collection elements regardless if they are missing or not. See example.

4

Hi there,

I have checked it and seems like a bug. Will check it and get back to you.

Tnx for reporting this.

Br,
Ivan

4 Likes
5

Hi there,

Just as an alternative approach, I suggest using Ungroup node before the GroupBy node and then use the List method in the GroupBy node. That will solve your issue for now.

:blush:

1 Like
6

Sadly this doesn’t work in my case because I have multiple sub-groups, meaning several columns containing to the same “sub-group” with same amount of elements in the list. One subgroup could have 2 elements and another 10. This leads to additional rows which then when grouping again leads to additional, wrong missing values. So I need to split the table up in these subgroups and then do the ungroup -> group and join again. Safe to say a fix for this issue would help me a lot.

7

This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.

9

Reopened and moved to Feedback & Ideas since @s.roughley asked about this. There is an existing ticket (AP-12262) but as far as I know it has not yet been addressed.

2 Likes
10

Thanks @ScottF

For public record this bug also applies to the Column Aggregator node (Unsurprisingly as I’m pretty sure they use the same AggregationOperators under the hood)

Steve

2 Likes