Start from a spreadsheet you already have

Most first databases aren't typed in — they're a spreadsheet somebody has been nursing for a year. The import wizard reads a CSV or Excel file, guesses what each column is, shows you the guess next to real sample values, and lets you correct it before anything is created.

The promise underneath: the wizard validates the entire file before it writes anything. There is no half-imported table to clean up afterwards.

Where Import lives

Three doors, one wizard:

• Starting fresh: click New database and pick the Import a spreadsheet card. Tabla makes the database first, then opens the wizard on arrival — and once the rows land it opens the database's name ready to type over, since a database made this way starts out called "Untitled database".
• Inside a database: the ⋮ button at the right end of the table tabs → Import data….
• Either way you land on a two-option dialog: Upload a CSV or Excel file, or Connect an Airtable base — that second path has its own guide.

Pick the file, read the guesses

Choose a .csv, .tsv, .xlsx or .xls file. The first row is read as headers; an Excel workbook imports its first sheet only, and the header line tells you how many rows were found and how many extra sheets were ignored.

For each column, the wizard shows the header, a few sample values, and its guess at a field type, worked out from the samples:

• true/false, 1/0, yes/no → Checkbox
• Numbers → Number — or Currency when the header sounds like money ("price", "cost", "total") and every value has two decimals
• 85% → Percent
• ISO dates and n/n/yyyy dates → Date
• Addresses and links → Email or Web address
• A column with few distinct short values ("Paid", "Pending") → Single select; delimited lists ("Fern; Indoor") → Multi select
• Anything long or multi-line → Long text; everything else → Single line text

One guard worth knowing: a column like 00123 is never guessed as a number, because a number would eat the leading zeros. IDs stay text.

The preview is editable

The guess is a pre-filled default, not a decision. Per column you choose:

• Create new field — rename it, or change the type from the dropdown. Blank headers become Column 3; duplicate headers get a 2 appended.
• Skip — the column never lands.
• Map to existing field — offered when you import into an existing table (that flow has its own page); only type-compatible fields are listed.

Slash dates get one extra control: Read dates as, month-first or day-first. 03/04/2026 is 4 March or 3 April depending on where the file came from, and when no sample has a day above 12 the wizard says so plainly — "both readings fit — worth a look." This is the one choice that can import silently wrong dates, so it's the one worth ten seconds.

Validate, then write

Click Import N rows. The importer coerces every cell of every row before creating anything. A value that doesn't fit its column — "TBD" in a currency column, a date the importer can't read — aborts the run with the row number and the reason, and nothing is created: no table, no fields, no records. Fix row 40,213, run it again.

When it finishes, you land on the table holding the rows you just imported — not on whatever table you happened to be looking at when you started.

Two honest edges:

• Appending to an existing table is different: rows are written in batches, and batches already committed stay. The error tells you exactly how many rows landed before the failure.
• A "select" column that turns out to have more than 100 distinct values is refused with the count — that isn't a select, and the wizard suggests importing it as text instead.

Limits

A CSV or TSV has no limit. Your browser reads the file where it already sits and streams the rows in, so a 500,000-row export imports like any other — with a progress bar and a Stop import button. The ceiling is your database's own 1,000,000 records.

An Excel workbook takes a file up to 20 MB and 50,000 rows per run, since a .xlsx is a zip and has to be opened whole. Save it as CSV to lift that entirely, or split the sheet. Everything around these is listed in Limits.

The questions people ask

What happens to formulas in my spreadsheet? Their current values arrive; the formulas themselves don't. A CSV never contained them anyway. Rebuild what still needs computing with a formula field.

Can I re-import the same file later? Yes, but nothing deduplicates — importing twice into one table means every row twice. Import into a new table when in doubt; deleting a table you got wrong takes one menu.

Does a Tabla export re-import cleanly? Yes — the export writes standard CSVs, and the importer reads its own exporter's output back without mangling values (phone numbers with a leading + included).