Table data

Double-clicking a table or view in the browser, or Open on its context menu, opens its page at the Data tab: its rows. The same table opens in the same page, and its structure and diagram are on the tabs beside the rows.

Filter and Order above the rows take what would follow WHERE and ORDER BY: city = 'Praha' AND joined > '2025', name COLLATE NOCASE. Enter in either reads the rows again.

A table’s rows can always be edited: the page finds each row by the table’s primary key, or, for a table that has none, by its rowid, which is then the first column shown. A view’s rows are read-only.

The rest of the toolbar works on the whole table: Populate… and Import…, below, and Empty…, which deletes every row after asking. F5 reads the rows again; with changes not saved yet, it asks first. Closing the page with changes not saved asks whether to save them.

Populating a table

Populate… fills a table with made-up rows, for trying out queries and indexes on more data than there is:

Each column gets one of:

Fill with	What goes in
Autonumber	the next number after the largest in the column
Autonumber from value	numbers counting on from the one given
Random number	a number of the given number of digits
Random text	letters, as many as the size says
Prefixed text	the given prefix followed by a number
Static value	the same value in every row
Date: now / random	a date as text, YYYY-MM-DD HH:MM:SS in UTC, as SQLite’s datetime() writes it
Date: … Unix time	the same, as seconds since 1970
Date: … Julian day	the same, as SQLite’s Julian day number
Ignore	nothing: the column’s default

Skip rows that break a constraint inserts with INSERT OR IGNORE, so a random value that collides with a UNIQUE one leaves that row out rather than failing them all. The rows go in together, in a savepoint: either all of them are added or none are.

Importing CSV

Import CSV… on a table adds rows to it from a file; on a database, or from the Database menu, it can make a new table for them:

File format

The delimiter — comma, semicolon, tab or pipe, guessed from the first lines — the quote, the encoding, lines to skip at the top, whether the first line names the columns, and what an unquoted field reads when it is NULL. A quoted field is never NULL. The encodings are UTF-8, UTF-16, ISO-8859-1 and -2, and Windows-1250, -1251 and -1252, on every system.

Into

An existing table, emptied first if wanted, or a new one, whose columns are named after the file’s first line and typed INTEGER, REAL or TEXT by what the first rows hold.

Columns

Which column of the file goes into which column of the table. Columns of the same name are matched by themselves; a column left out takes its default.

Preview shows the first rows as they will be inserted.

The import is one transaction — or, inside a transaction already open, one savepoint — so a row that fails, or Stop, leaves the table as it was.