Different names for the same name ID: need to use name IDs only when grouping

[lmullen]

In elections, NNV sometimes reports votes separately for people with similar names WHO HAVE THE SAME NAME ID. In effect they are treating them like they are different people, but they are likely the same person. For instance, Lucas Elmendorf and Lucas Elmendorf Jr. both get columns in this election. But they are the same person with the same name ID.

So when we are doing grouping, we have to rely on the name ID only, and disregard the names altogether. But we will have to check that everyone has a name ID. Anyone without a name ID (e.g., null) will have to have their name as a name ID, or we will have to add them.

[lmullen]

An example election: http://elections.lib.tufts.edu/catalog/tufts:ny.uscongress4.1800

[lmullen]

There are approximately 3,300+ distinct names without name IDs. There are a lot of people with just last names. But there are some like Lucas E. Elmendorf and Lucas Elmendorf who are undoubtedly the same person as the Lucas Elmendorf and the Lucas Elmendorf Jr who do have name IDs.

[lmullen]

There are 3,765 name ID which have duplicate names, some of them with as many as 10 variant names.