Tables in KNIME have what’s called a “table specification”. You can see this by going to the Spec tab in a tabular output port, like this:
You can see at a glance that I have 248 rows and 5 columns. Also listed are the ranges of each numeric variable, and the possible values shown for string variable (Cluster_0 through Cluster_3). Based on what we see here the number of possible values for the Cluster Memberships is 4.
Now it happens that you could have cases where you have a very large number of possible values in a particular string field - so many that it might cause problems with downstream nodes, where there are limits on the number of categories that can be reasonably processed, for example (some of these are discussed in the linked threads above).
This is where the “restrict number of possible values” option in the Domain Calculator node comes in - you can set a hard limit so that other nodes don’t run into trouble. The Domain Calculator has other uses too, mentioned in its node description - things like recalculating the bounds of numeric variables if those have changed (maybe in a loop or something).
Does that help clarify?