Scenarios: repeat a table clean-up on the next file

The same export arrives every month — a customer list, a ledger, a report from another system — and you do the same things to it: tidy the names, add a column, mask the IDs, keep the rows you need and save it to Excel. A scenario remembers those steps once. After that the app repeats them: on the open table, on a whole folder or from the command line. Below are six steps on a real example.

Updated 2026-09-30 · format CSV

The example

A monthly export in DBF or CSV gets four things done to it before it goes on:

1. last, first and middle names — upper case and without stray spaces, as the receiving system requires;
2. a single “Full name” column built from the three;
3. ID numbers replaced with sequence numbers so the file can go to a contractor;
4. only rows with an amount of 3,000 or more, sorted by name.

By hand that is ten minutes and four dialogs. As a scenario it is one click.

Step 1. Process the file once, as usual

Open the export and do what you do every month with the usual commands: “Edit” → “Clean up”, “Transform” → “Calculated column”, “Anonymize columns”, a filter. There is nothing to record: the tab keeps a log of operations by itself, and Ctrl+Z removes a step from the log too.

When the result looks right, open “Transform” → “Save steps as scenario…”.

The save dialog lists the recorded steps with check boxes — untick what you don’t need. It also shows what is saved besides the steps (the encoding if you chose it, the delimiter, the sheet, the final filter and column order) and what is not, because it can’t be carried over to another file: edits of single cells, replacing in selected rows only, rows added by hand.

Step 2. Find the scenario in the list

All scenarios are in the app menu (☰ at the top left) → “Scenarios…”. Each one shows its steps in words, the file it was recorded on and buttons to apply it to the open table, open it in the editor, run it on a folder or export it to a file.

If the open table lacks columns the scenario needs, the list says so before anything runs.

Step 3. Open the scenario in the editor and check it on a sample

“Open” shows the whole scenario. On the left is the chain: input (how to open the file and which file to use as a sample), the steps in order, the result (filter, sort, result columns) and the output (format). On the right are the settings of the selected step and a sample table after it.

The sample is computed for real, on your file, by the same code that runs later. Each card says what the step did to the sample: how many cells changed, which columns appeared, how many rows were kept. A step can be switched off — it is skipped but stays in the scenario. Drag steps or use Alt+↑/↓ to reorder them. The left panel can be widened by dragging the border; the width is remembered.

Step 4. Add a step or adjust one

Hover over the line between steps and a “＋” appears. It opens a catalog: rows (filter, sort, duplicates), values (replace, clean up, fill, anonymize), columns (formula, add, rename, split, merge, delete), checks and lookups.

Every step is edited with a form. Columns are picked from a list — the ones that exist at that point of the chain, including those created by steps above. A value that changes from run to run, such as a month or a threshold, can be written as {{Month}}: the scenario asks for it each time it runs.

Step 5. Check that the scenario fits another export

The usual trouble with repeated processing is that the next file is slightly different: the supplier renamed a column, the export came from another program. Choose such a file as the sample on the Input card — the step that would fail turns yellow before anything runs.

Steps below the error are marked “not checked”. Fix the step — pick another column or switch it off — and the sample is recomputed at once. The app never skips a step silently: a wrong result nobody noticed is worse than a stop.

Step 6. Run it on a folder — or from the command line

“Run…” opens batch conversion: pick a folder or files, and each one is opened with the scenario’s settings, goes through its steps and is saved in the output format — as one ZIP or as files in a folder. A file the scenario stopped on goes to the error list with the step number; the others are exported.

To apply a scenario to one open table, “Transform” → “Apply …” is enough. Steps that change values switch editing on: the file on disk changes only when you click “Save”, and everything can be undone with Ctrl+Z.

The Windows app runs a scenario without a window too — for example on a schedule from Task Scheduler. Export the scenario to a file first:

Tabulens.exe convert C:\exports\*.dbf --to xlsx -o C:\done --recipe monthly.recipe.json --param Month=2026-10

Sharing a scenario

A scenario is a small .recipe.json file. Send it by email and the recipient loads it with “Load from file…” in the Scenarios window. It contains no table data — only column names, conditions and formulas. Before saving, the app lists any specific values written into conditions or replacements (an ID number in a filter would travel with the file) and offers to embed the lookup tables the scenario uses, so the recipient doesn’t have to find them.

What a scenario includes

• Included: filter and sort, calculated columns, lookups, value checks, find and replace, clean-up, fill, anonymization, adding, renaming, splitting, merging and deleting columns, marking duplicates; the file’s open settings; the final column order and visibility.
• Not included: edits of single cells, replacing in selected rows only, rows added by hand, merging with other files.
• Columns by name, not by position: a scenario fits any file with the same columns in any order. Letter case doesn’t matter — DBF stores names in upper case.
• Steps that change values need a format that can be edited: DBF, CSV, Excel (XLSX), ODS, JSON, XML, SQLite. Filters, formulas, lookups and checks work with any format.

FAQ

Free to use. Your file is not uploaded to a server. Windows version — 5.4 MB, no installation: details. Pro — $29 once: price.

Related