Convert dBase to DuckDB

Working from a dBase file

dBase III and IV gave the format its five original field types, and that small vocabulary is why a dBase table converts so cleanly: there is very little in one that a modern database cannot hold exactly.

• →Character, Numeric, Date, Logical and Memo are the whole type system. A Numeric field declared with no decimals becomes a 64-bit integer; one with decimals becomes an exact decimal of the declared width and scale.
• →dBase has no time-of-day type at all, so a Date field converts to a pure date and never to a timestamp at midnight.
• →dBase IV memo blocks carry an 8-byte header and dBase III's do not, so a .dbt written by either generation is read with the block layout it actually uses.

Decimals and text in the .duckdb file

DuckDB takes the source's declared types more nearly unchanged than any other target here, which makes the two places where it cannot the things worth reading: the width of a decimal, and the length of a string.

• →DECIMAL tops out at 38 digits, so a wider source column is clamped rather than refused: an Oracle NUMBER(50,10) is created here as DECIMAL(38,10), and the scale is never allowed to grow past the precision holding it.
• →A decimal that declared no scale gets six places for its fraction rather than none, inside that same rule: a source precision of 10 becomes DECIMAL(10,6), one of 5 becomes DECIMAL(5,5), and one that declared no precision either becomes DECIMAL(18,6).
• →Strings become VARCHAR with no length attached to them. DuckDB's VARCHAR is unbounded, so no column has to be sized in advance and no row can be turned away for being longer than the column it goes in.

Reading dBase

• →Numeric fields with decimals are read as exact decimals, not floats, so currency does not drift.
• →Character field widths are read from the header and carried into the target.
• →The code page declared in the file header is used, so Cyrillic, Western European and other legacy DOS encodings decode correctly instead of turning into mojibake. 157 code pages are recognised.
• →Files written by dBase III or Clipper declare no code page at all — the header byte for it did not exist yet. For those, the most plausible encoding is chosen by inspecting the text, and the conversion tells you it had to guess. The API takes a sourceEncoding parameter (e.g. cp866) to force it instead of guessing, though the widget above does not have a control for it yet.
• →Memo columns keep their text in a separate .dbt or .fpt file. If the .dbf declares one and it was not uploaded, the conversion stops and tells you which file to add, rather than handing you a table whose long text is silently blank.

Writing DuckDB

• →A real .duckdb file, not a script — open it straight away with the DuckDB CLI.
• →DuckDB has genuine DECIMAL and BOOLEAN types, so nothing is flattened on the way in.

Convert dBase to something else

• MySQL
• PostgreSQL
• Oracle
• SQL Server
• SQLite
• CSV
• Excel

Convert something else to DuckDB

• CSV
• Excel
• SQLite
• SQL dump
• JSON
• XML
• Parquet
• MS Access
• PostgreSQL backup