Control / params · All operations

Change an engine setting

A variant of Set parameters

Change, restore or list the configuration options of the engine or session, such as the default schema, an extension or a seed.

Status

In arcform

Not built yet

arcform does not do this yet; it would live in how a Protocol runs. Its place on the roadmap

What other tools call it

Meridian's classification; an independent blind rating agreed on 95.7% of 324 rows (variant, version 3).

Apache DataFusion (1)
  • SHOW ALLread 2026-09-27

    List the session configuration options

DuckDB (4)
  • LOAD / INSTALLread 2026-09-27

    Install and load an extension

  • SET / RESETread 2026-09-27

    Change or restore a configuration option

  • USEread 2026-09-27

    Choose the default database and schema

  • duckdb_settingsread 2026-09-27

    duckdb_settings table function

Polars (4)
  • Extension type registry (family)read 2026-09-27

    3 functions: register_extension_type, unregister_extension_type, get_extension_type

  • Scan and sink options (family)read 2026-09-27

    3 classes: PartitionBy, FileProviderArgs, ScanCastOptions; partitioned output files and casts applied on scan

  • StringCache (family)read 2026-09-27

    4 entries: StringCache, enable_string_cache, disable_string_cache, using_string_cache; documented as doing nothing

  • set_random_seedread 2026-09-27

    Set the global random seed

subtoken (2)
Talend (1)
  • tLibraryLoadread 2026-09-27

    Load Java libraries for use in the Job

VisiData (6)
  • edit-optionread 2026-09-23

    Edit the option in current Options Sheet row

  • open-configread 2026-09-23

    Open the config file as a text sheet

  • options-globalread 2026-09-23

    Open Options Sheet to edit global options

  • options-sheetread 2026-09-23

    Open Options Sheet to edit options for current sheet

  • unset-optionread 2026-09-23

    Remove the option override in current Options Sheet row

  • visidatarc (saver)read 2026-09-23

    Save as Options Sheet as a .visidatarc config file

Background

DuckDB has configuration options that change the system's behaviour, set with the SET or PRAGMA statement and reset to their original values with RESET.[1] In DuckDB, the value of a configuration option can be queried with the current_setting() function or the duckdb_settings() table function.[1] Apache DataFusion shows the current session's configuration options with the SHOW ALL command or the information_schema.df_settings view.[2]

When to use it

In DuckDB, a GLOBAL configuration value is used across the entire DuckDB instance, while a SESSION value is used only for the current session.[3] In DuckDB, GLOBAL settings are not written to the database file, so a new instance uses the defaults unless the options are set again.[3] In DuckDB 2.0, the TimeZone option defaults to the SESSION scope, so setting it changes the time zone for the current session only.[3] Set the options a result depends on, such as the time zone or the default sort order, in the pipeline itself, not in a client's startup file that another machine will not have. (Meridian's view)

A pseudorandom number generator's sequence is completely determined by its seed, so reinitialising it with the same seed produces the same sequence of numbers.[4] In Polars, set_random_seed sets the global random seed, which determines things such as shuffle ordering.[5] Fix the seed in any run that samples or shuffles, so the run can be repeated exactly. (Meridian's view)

The method

In DuckDB, the USE statement sets a default database, schema, or database and schema, so a table created without a fully qualified name is created in the default database.[6] In DuckDB, the INSTALL statement downloads an extension, and the LOAD statement loads an installed extension into the current session.[7] In DuckDB, LOAD fails with an error if the extension is not installed yet, while INSTALL AND LOAD, introduced in DuckDB 2.0, does both in one step.[7] When a DuckDB query uses a function or feature of a known core extension, autoloading installs and loads that extension automatically.[7] Name each extension a pipeline needs in an explicit step rather than relying on autoloading, so a reader can see the dependency. (Meridian's view)

In DuckDB 2.0, duckdb_settings() returns only regular settings by default, and passing include_all := true returns every setting, debug-only and deprecated ones included.[8]

Compared with the alternatives

A configuration file stores data used to configure a software system, such as an application, a server or an operating system.[9] Some programs read their configuration files only at startup, while others periodically check the files for changes.[9] A setting changed by a step travels with the pipeline, while one kept in a configuration file stays with the machine. (Meridian's view)

Sources

  1. [1]Configuration – DuckDBdocumentation
  2. [2]Information Schema — Apache DataFusion main documentationdocumentation
  3. [3]SET and RESET Statements – DuckDBdocumentation
  4. [4]Random seedsecondary source
  5. [5]polars.set_random_seed — Polars documentationdocumentation
  6. [6]USE Statement – DuckDBdocumentation
  7. [7]LOAD / INSTALL Statements – DuckDBdocumentation
  8. [8]DuckDB_% Metadata Functions – DuckDBdocumentation
  9. [9]Configuration filesecondary source