apache / apache/datafusion

[EPIC] Support ANSI mode

Open
#23,929 6 comments 0 reactions 0 assignees View on GitHub
enhancement spark
Dominant language
Rust
Stars
9.3k
Forks
2.4k
Avg merge
3d 7h
Merged PRs (30d)
344

Description

### Is your feature request related to a problem or challenge?

## Description

Apache DataFusion currently follows permissive SQL semantics for many runtime errors (for example, invalid casts, arithmetic overflow, division by zero, and malformed input). To improve SQL compatibility and make behavior configurable, DataFusion should support an ANSI SQL evaluation mode.

ANSI mode would cause operations that encounter invalid input or runtime errors to return an error instead of silently producing `NULL` or another permissive result. This behavior is expected by users migrating from systems that support ANSI SQL semantics (such as Spark SQL with ANSI mode enabled) and enables stricter correctness guarantees.

Supporting ANSI mode should provide:

* A configurable session-level ANSI mode.
* Consistent error semantics across expressions and built-in functions.
* Preservation of existing permissive behavior when ANSI mode is disabled.

This issue tracks adding the infrastructure and expression/function changes required to support ANSI mode throughout DataFusion.

The initial effort started with adding DataFusion parameter https://github.com/apache/datafusion/pull/18635 which is false by default

```
pub enable_ansi_mode: bool, default = false
```

### Describe the solution you'd like

_No response_

### Describe alternatives you've considered

_No response_

### Additional context

_No response_

Contributor guide

Open the contributing guide

Research direction

The issue names no files or tests. Start by reviewing the session parameter introduced in PR #18635 and map the expression and function areas affected by invalid casts, overflow, division by zero, and malformed input. Done means configurable session-level behavior, consistent errors when enabled, and unchanged permissive behavior when disabled.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust, sql
Domain
databases
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
28/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.