Where your files are saved

This section answers the question people ask most about DataTray, so it is worth reading once properly.

Never overwrite, never in place

DataTray never modifies your source file and never replaces an existing output. Every tile writes a new file. If you clean orders.csv, orders.csv is exactly as it was when the run finishes — byte for byte, with the same modified date.

Every utility writes into a shared output folder, inside its own subfolder:

Change the root folder under Settings ▸ General ▸ Output folder with Browse…, or open it in File Explorer with Reveal.

Writing beside the original was considered and rejected. Almost every DataTray operation keeps the file's extension, so cleaning orders.csv "alongside" could only ever produce orders (1).csv — two identically named files and no way to tell which is which.

Names come from a template, not from a numeric suffix

Output names are built from a small token template, so the name says what happened and when:

The available tokens are {stem} (the source filename without its extension) and the date and time tokens {YYYY} {YY} {MM} {DD} {HH} {mm} {ss} {Year} {Month}.

That is why running the same recipe every morning produces one clearly labelled file per day rather than a pile of (1), (2), (3) with no way to tell them apart. A numeric suffix is the fallback, not the norm: it only appears when you run the same operation on the same file twice on the same day.

A companion file always accompanies a lossy run

Whenever an operation dropped or rejected something, what did not make it is written beside the output:

This is the whole argument against the tools people use today: Excel's Remove Duplicates tells you a number and nothing else. DataTray hands you the rows. Nothing is lost silently.

Reports

Several tiles can write a report — a single self-contained HTML file you can open in a browser, mail to a colleague, or print to PDF, with an Excel version available where the numbers are more useful in a grid. Reports land in the same subfolder as the output they describe, named <source> - <report kind> - <date>.

If something goes wrong here

• I cannot find my output file. Open the job in the Job Queue and press Reveal. That is always faster than hunting through folders.
• My original is unchanged. That is the guarantee, not a bug. The result is a new file in the tile's output subfolder.
• There is an extra file next to my result. That is the companion file described above. Open it — it is telling you what was dropped.