I was wondering if it would be possible to add an option to exclude counting missing or even count the missing values only when using the Collection Size node?
The use case is, when creating a List or set with missing values to ensure after ungrouping all ends up in the right spot again, I want to quickly check without ungrouping if there are any missing values and then maybe discard them all together / reprocess.
Iâve added both the requested options, and also an additional option to only count the unique (or âdistinctâ in SQL or Java Streams nomenclature) values (for a set column thereâs no effect with this, as they are always unique, but for a list, it saves going via an intermediate set)
Existing nodes will acquire the new options, but keep their behavior (i.e. the new options will be populated as shown, and the output count columns will be âintâ types). New nodes added to workflows will also default to these settings, as they seem to be the most obvious defaults, but the output count columns will be âlongâ types
If that all looks ok to you, then I will try to get the public version updated in the coming days?
It should now be available in the nightly build/trunk channels, and the stable channels for KNIME 5.3.x and KNIME 5.2.x. Iâm hoping to get it out in KNIME 5.1.x and 4.7.x too, but there is an issue I need more help resolving before that can happen.
Thanks again for the suggestion - definitely one of those âWhy didnât we think of that sooner?â ones!
Some sort of help I might be able to provide giving back something maybe?
About
âWhy didnât we think of that sooner?â
Did you, as the de facto leader in collection nodes, thought about doing math with collections containing numerical values similar to âmatrix magicâ? Bonus for nerds, doing that using the GPU