Developing a merge routine

I find it helpful to think of merging data as the parking lot of the data cleaning process. Merging data and parking may seem routine, but they are where a disproportionate number of accidents occur. One approach to getting in and out of parking lots without an incident occurring is to use a similar strategy each time you go to a particular lot. It could be that you always go to a relatively low-traffic area and you get to that area the same way most of the time.

I think a similar approach can be applied to getting in and out of merges with our data relatively unscathed. If we choose a general approach that works for us 80 to 90 percent of the time, we can focus on what is most important—the data, rather than the techniques for manipulating that data.

In this recipe, I will demonstrate the general approach that works for me, but the particular techniques I will use are not very important. I think it is just helpful to have an approach that...