The question of one file versus three isn't really about storage space, it's about how you work. If you're mostly pulling data for a single category at a time, splitting the files is the more honest reflection of your workflow. A dashboard that only needs category A shouldn't have to load categories B and C just to get there. That's not a technical opinion; it's a matter of reducing the amount of data your system has to touch on every request. The user's own habit of focusing on one category at a time is the strongest signal here, and it points toward separate files.
The conflicting advice from the two tools isn't a contradiction, it's a clue that the "right" answer depends on what you're optimizing for. One file can be more efficient for storage because it avoids duplication and keeps related records together. But that efficiency comes at a cost: every time you open the dashboard, you're loading the entire dataset, even the parts you don't need. For large datasets, that translates into slower response times and higher memory usage. The tools that said "one file" were likely thinking about disk space. The one that said "split" was probably thinking about retrieval speed. Both are valid, but only one matches how you actually work.
On format, the user's mention of Excel and CSV is worth addressing directly. CSV is the better default for large datasets because it's plain text, which means faster reads and no hidden metadata bloating the file. Excel adds formatting and formulas, but those come with overhead, especially when the file grows large. If the user is pulling data into a dashboard, CSV is often the more practical choice because it's lightweight and universally readable. Excel has its place, but for this kind of structured, category-based data, CSV aligns better with the goal of speed and simplicity.
So here's the practical takeaway: split the files, use CSV, and let your dashboard point to the category you need at any given moment. You'll lose a little storage efficiency, but you'll gain responsiveness and reduce the load on your system every time you refresh. And if you ever need all three categories together, you can always combine them on demand, that's a much lighter operation than constantly filtering a massive single file. The tools gave you opposite answers because they were optimizing for different things. You know what you need. Choose the approach that matches your workflow, not the one that sounds more impressive in theory.