Skip to content

Marivo 0.5.5

Marivo 0.5.5 adds a built-in local datasource and clarifies terminal SQL and typed analysis boundaries.

  • The built-in default datasource provides credential-free, in-memory DuckDB for CSV, Parquet, and JSON sources without configuration or registration. It appears in md.list() and md.describe("default"), uses independent non-persistent connections, and reserves the name default.
  • md.raw_sql(...) now executes the caller’s SQL unchanged and returns every row from that query. Query scope is explicit in SQL, making terminal source exploration predictable.
  • Supported analytical calculations remain in the typed flow after observation, while presentation reads, unsupported methods, and source SQL have distinct Help and repair guidance.
  • md.raw_sql() no longer applies a client-side row or byte cap. Control query size with filters, partition predicates, aggregation, and SQL LIMIT; all returned rows load into client memory, and LIMIT does not bound source scan cost.
  • Raw SQL is trimmed only for whitespace and trailing semicolons before execution. Empty, comment-only, and semicolon-only input is rejected before a connection is acquired. Read-only execution depends on backend protections and database permissions; Trino relies on database-side permissions.
  • Datasource and semantic catalogs, Help, doctor checks, and packaged agent guidance now describe the built-in datasource and the caller-owned SQL scope.
  • This release removes the limit parameter from md.raw_sql(). RawSqlResult no longer exposes requested_limit or is_truncated; update callers to express result scope in the SQL statement and treat the result as complete for that query.
  • Updated the semantic-layer, analysis-workflow, first-analysis, datasource, Help, and packaged agent guidance in English and Chinese for local file sources, terminal SQL, typed execution, and repair routing.
Terminal window
curl -fsSL https://marivo.io/install.sh | bash

In mainland China, use:

Terminal window
curl -fsSL https://marivo.io/install-marivo-cn.sh | bash

For manual installation:

Terminal window
pip install marivo

Marivo 0.5.5 requires Python 3.10 or newer.