Doing one-to-many merges
In one-to-many merges, there are unduplicated values for the merge-by column or columns on the left data table and duplicated values for those columns on the right data table. For these merges, we usually do either an inner join or a left join. Which of those two join types we use matters when merge-by values are missing on the right data table. When performing a left join, all the rows that would be returned from an inner join will be returned, plus one row for each merge-by value present on the left dataset, but not the right. For those additional rows, values for all the columns on the right dataset will be missing in the resulting merged data. This relatively straightforward fact ends up mattering a fair bit and should be thought through carefully before you code a one-to-many merge.
This is where I start to get nervous, and where I think it makes sense to be a little nervous. When I do workshops on data cleaning, I pause before starting this topic...
