Fix SQLite CSV Import Errors: Types, Quotes and Columns
CSV has no reliable built-in column types. Inspect the preview before importing and keep identifiers as TEXT unless you intend numeric conversion.
Check delimiter and header first
A single giant preview column often means the wrong delimiter. Select comma, tab or the delimiter used by the file. If the first data row appears as column names, disable the header option. Every nonempty record must match the expected number of fields.
Keep identifiers as text
Postcodes, account identifiers and values such as 00123 should normally be TEXT. SQLite Lab defaults to TEXT. Choose INTEGER or REAL only for actual numbers; currency symbols, thousands separators and invalid numeric strings can make an explicitly numeric import fail. Empty numeric fields become NULL.
Repair quoting at the source
A value containing the delimiter or a newline needs CSV quoting. A double quote inside a quoted field is represented by two double quotes. Export UTF-8 when possible. Do not fix a broken file by removing every quote or splitting lines manually; that can damage multiline cells.
Retry without partial tables
The importer rolls back its new table when a row or numeric conversion fails. Correct the source or import configuration and try again with a non-conflicting table name. After import, inspect row counts and representative values before exporting the database. Use the CSV converter page to open the import preview.
Keep exploring
SQLite syntax reference · Practice with a learning path · Sample datasets
CSV Data Types: Preserve IDs and Clean Numbers
Reference: Official SQLite documentation