Import and export

Bring existing spreadsheets into a table without retyping anything, and take your data back out whenever you need to share it or open it in another tool. This page covers the import dialog step by step — choosing a file, creating a new table or appending rows, and how column mapping and type conversion behave — plus every export format and how to copy a table straight to the clipboard.

Import a file

Import reads CSV and spreadsheet files (XLSX, XLSM, XLSB and the older XLS). You start it from the Import CSV / Excel button on the Bases overview; a dialog opens with a preview so you can check what will be imported before anything is written to a table.

1Click Import CSV / Excel and choose the file. The dialog reads it and shows how many columns and rows were detected, with a preview of the first rows.
2Choose the Destination: Create new table from the file, or Add to current table to append the rows to the table that is open.
3If you chose a new table, confirm or edit the New table name (it starts from the file name).
4Review the preview — the header row becomes the column names, and the rows below become the data.
5Confirm with Import; the rows appear in the table.

The file needs a header row plus at least one row of data. Completely empty lines are ignored. Only the first sheet of a spreadsheet is read, so put the data you want on the first tab.

When you create a new table, every column comes in as plain text — types are not guessed at this stage. After importing, open the Structure editor and set the real type (number, date, select, and so on). See "Column types".

Append to an existing table

Add to current table appends the file's rows to a table you already have. Columns are matched by header name — a file column lines up with a table column when their names match (ignoring case and surrounding spaces). This is why headers matter.

File columns whose name matches a table column are imported into that column.
File columns with no matching table column are ignored — nothing new is created on append.
Table columns with no matching file column are simply left empty in the new rows.
Values are converted to fit the target column type, so the imported data lands in the right format instead of as loose text.

Conversion is forgiving: number and currency columns accept values like "R$ 1,234.50", checkboxes read true/yes/1/x, multi-value columns split on commas, ratings are clamped to 0–5 and progress to 0–100. Reference and task columns match by the visible label or title; when there is no match the cell is left blank. Formula, lookup, rollup, and reverse-link columns are never imported — they recalculate on their own.

Append is only available when a table is open, since it needs a destination. With no table open, the dialog falls back to creating a new one.

If the connection drops mid-import

Large imports are sent to the server in batches. If a batch fails because of a hiccup in the network, it is retried automatically a few times; if the connection is really down, the pending rows are queued safely on your device and sent as soon as you are back online.

When that happens you see the notice "Changes waiting to sync" — nothing is lost, and re-sending is safe: rows are written by identity, so a retried batch never duplicates records.

Export

Export downloads the current table as a file. Pick the format from the Export menu above the table (available in the record views — Table, Gallery, Kanban, List, Calendar).

CSV — universal spreadsheet format, opens in Excel, Sheets and almost anything (saved with a mark that keeps accents correct in Excel).
Excel (.xls) — a spreadsheet file that opens directly in Excel with the header styled.
JSON — one object per row, keyed by column name, with real values (numbers as numbers); reference and task cells come as their visible label. For developers and other tools.
Markdown — a Markdown table you can paste into notes or documentation.

To copy instead of downloading, use Copy table in the table's menu (the ••• next to its name). The whole table is copied as tab-separated text, so you can paste it straight into Excel or Google Sheets. If the browser blocks clipboard access, nothing is copied and you are told.

What carries over

File exports write what you see: reference and task cells come out as their visible label, checkboxes as true/false, and multi-value or attachment cells as a comma-separated list. Rows are exported in the order they appear. To learn how the cells themselves work, see "Rows".

Bases is also the foundation of the Forms module: a public form is built from a table and each submission becomes a new row that you read back here — so a form is another way to feed rows into a table without importing a file.