Color nodes by partition incorrectly assigns attribute values for items with empty values

[serenalotreck]

Expected Behavior

If there are nodes in a network that don't have a value for an attribute in the Data Laboratory (i.e. an empty string fills the cell), they should come up as their own group when coloring nodes by partition for that attribute

Current Behavior

Gephi randomly assigns nodes with no value to one of the other groups, with no indication of how or why. It's very problematic because if there aren't a ton of missing values immediately evident in the Data Laboratory, it's not obvious that the representation of your data is completely incorrect for those values; I assumed that all of my nodes had been assigned a continent and only noticed because I was looking for something completely unrelated.

Possible Solution

Steps to Reproduce

1. Make a graph with some attribute, in my case it's called continent
2. Delete the value from one or more cells in the Data Laboratory under the continent column
3. Go to Overview, and under the Appearance tab, select Nodes, Color, and Partition, and color by continent
4. There will be no additional category for the nodes with the deleted values, and they will have been randomly assigned to another category

Context

I'm visualizing a graph that I made in networkx. I added attributes from a dictionary to the whole graph and saved it as .graphml; however, I did not explicitly add NaN values for those nodes that did not get assigned to a continent. I assumed that Gephi would be able to figure out that an empty attribute didn't belong in one of the other categories.

Your Environment

• Version used: Gephi 0.10.1
• Operating System: Windows 10

[serenalotreck]

I had to do some manual arranging of the graph layout, so I'd really like a way to fill in NaN values for the empty cells in Data Laboratory as a temporary solution; but I can't figure out how to do it with the options I see. Would appreciate if anyone had a solution for that.