Skip to content

data-types.md

Latest commit

 

History

History
32 lines (24 loc) · 1.33 KB

data-types.md

File metadata and controls

32 lines (24 loc) · 1.33 KB

Data types & NULL handling

The row model

The fetch methods return plain Python containers — there is no Record type:

  • fetch(...) → list[dict] (one dict per row, keyed by column name)
  • fetchall(...) → list[list] (one list per row)
  • fetchval(...) → a single scalar value

If you need a richer row object, wrap the results yourself.

NULL in BOOLEAN/DATETIME columns depends on the server version

A 2026 dqlite server change (upstream commit f30fc99, "query: preserve SQLITE_NULL type for NULL values") changed how a NULL cell is encoded in columns declared BOOLEAN, DATE, DATETIME, or TIMESTAMP:

  • Before that change: a NULL was sent with the column's coerced type — BOOLEAN with value 0 (decodes to False), or an empty ISO-8601 string (decodes to ""). Indistinguishable on the wire from a real False / empty string.
  • After that change: a NULL is sent as SQLite NULL and decodes to None.

This client decodes faithfully whatever the server sends — the server version drives the behavior, and there is no handshake field that distinguishes the two. So code like if row["flag"] is None: will silently miss rows against an old cluster, then start matching after the cluster is upgraded. Check your dqlite cluster version before relying on is None for BOOLEAN / DATETIME columns.