The question this reader is asking is the right one, and the instinct to reach for Power Query is sound, but the real issue here is not the tool, it's the mental model. What they are describing is a classic data fusion problem, and the accurate term they are searching for is a "time-series join" or "as-of merge." That is the phrase that will unlock hours of frustrated forum scrolling and turn a vague sense of "I should probably pivot this somehow" into a concrete, searchable solution. Power Query can absolutely do this, but only once you stop thinking about matching rows and start thinking about matching timestamps.
The practical reality is that your sensor data is not messy because it is wrong, it is messy because it is asynchronous. Each device operates on its own rhythm, and that is not a flaw in your setup; it is the nature of the problem. When one sensor logs every minute, another every hour, and a third only when an event fires, you are not dealing with a spreadsheet issue. You are dealing with a temporal alignment issue. The blank cells you are worried about are not failures, they are the honest representation of what the physical world actually measured at that moment. The goal is not to force every sensor into a perfect grid, but to create a single timeline where each timestamp carries the best available information from every source, even if that means carrying a lot of empty space.
The practical path forward is to stop trying to make the data look pretty and start building a scaffold. In Power Query, you would bring in each CSV as a separate query, then use a full outer join on the timestamp column. That is the key move, not an inner join, which would drop all the moments where only one sensor reported, and not a merge that assumes a perfect one-to-one match. A full outer join preserves every timestamp from every sensor, and then you simply let the blanks be blanks. The result will be wide, sparse, and completely usable for analysis. You can then either leave it as is or use a fill-down or interpolation step if you want to estimate values between readings, but that is a choice, not a requirement.
What this reader is really doing is taking the first step toward a more mature relationship with their data. They are not asking for a magic button; they are asking for the vocabulary to understand the problem. That is the mindset that separates people who fight spreadsheets from people who build systems. So here is the concrete takeaway: stop searching for "combine CSV files by time" and start searching for "Power Query full outer join on datetime." That single phrase will get you further than any tutorial on merging tables. Your master sheet is not a destination, it is a byproduct of asking the right structural question. And once you ask it, the answer becomes almost obvious.