[DISCUSSION] DataFusion Road Map: Q1 2026

[alamb]

Introduction

DataFusion is a community project, and unlike many other open source projects there is no "DataFusion Company" that pays contributors and determines their priority. The project's evolution and priorities is determined by the community members able and willing to devote their time and energy to driving the project forward (PRs, code review, planning, etc)

As we have done in the past, I think it is helpful to have a public discussion about what features people are planning to work on to:

1. Find others interested in the same features (to combine efforts and collaborate)
2. Find areas of broad community interest (areas that many users want are more likely to get the requisite focus to get done)

Prior Art

Per the Roadmap Section in the docs:

• The current list of EPICs can be found here.
• The current list of PROPOSAL EPIC (that are not yet underway) can be found here.

Epics offer a high level roadmap of what the DataFusion community is thinking about.

Community Request

1. Please create or leave comments on proposal epics you are interested in helping with (to help drive / build momentum). Commenting on this ticket is also great if you are looking to help out.
2. Comment below about what you are planning to help with / work on

Context

Previous discussions:

• [DISCUSSION] DataFusion Road Map: Q3-Q4 2025 #15878
• [DISCUSSION] 2025 Q1-Q2 Roadmap #14580
• [DISCUSSION] 2024 Q4 / 2025 Q1 Roadmap #13274

[alamb]

I personally plan to focus in Q1 on performance

1. [EPIC] Parquet filter pushdown into scan #3462
2. [EPIC] ListingTable object store usage improvements #17214
3. [EPIC] Make DataFusion the top of the ClickBench Parquet leaderboard #18489
4. [EPIC] Support VARIANT type for unstructured data #16116 (and related technical areas like expr pushdown)
5. Make adding an extension type a pleasant experience (epic TBF, see here)

[comphead]

There was also a demand of supporting lambda syntax like https://duckdb.org/docs/stable/sql/functions/list#list_filterlist-lambdax to unblock mostly array related builtin functions, not sure if it fits in Q1, but I probably try to invest some time in there.

[Omega359]

There was also a demand of supporting lambda syntax like https://duckdb.org/docs/stable/sql/functions/list#list_filterlist-lambdax to unblock mostly array related builtin functions, not sure if it fits in Q1, but I probably try to invest some time in there.

A somewhat similar lambda PR timed out recently - #17220

[Omega359]

What I'd like to see worked on and what I myself am capable of working on are unfortunately not aligned :/ That being said I myself am likely to

• take another stab at logical planning performance issues Planning time for queries with many columns with union and order by is very slow #17261
• perhaps look into what it would take to implement something similar to Spark's flatMapGroups
• additional improvements for timestamp with tz functionality datafusion.execution.time_zone is not used for basic time zone inference #13212 (comment)