Dropped Or Kept: The Honest Choice About Unmatched Rows

Records with no partner on the other side either vanish or stay with gaps, and choosing carelessly rewrites your findings.

When you line two collections up, some records will find no partner. A customer who has never ordered. An order whose customer record was removed. You have a genuine choice about what happens to these, and it is a choice with consequences. The default behaviour in most people's mental model is that unmatched records simply disappear from the result, and often that is what happens. Your list of customers with their orders quietly becomes a list of customers who have ordered, and the ones who never did are gone, silently, without appearing anywhere.

Sometimes that is precisely what you want. If your question was about purchasing behaviour, customers who have never purchased may be genuinely irrelevant. But notice how easily it goes wrong. If you were asked how many customers we have and you answered from a combined list, you have answered a different question: how many customers we have who also have orders. The number will be smaller and entirely believable. Nobody will query it. This is the single most under noticed way that a report ends up understating a population.

The alternative is to keep every record from one side regardless of whether it found a partner, leaving the unmatched ones with empty spaces where the partner information would have been. That preserves your population honestly, but it hands you a new responsibility: those empty spaces are absent values, and absent values behave in ways that will surprise you when you start counting and averaging. There is no free option here. Either you drop records or you carry gaps. Decide deliberately, write down which you chose, and know what it does to the numbers that follow.