DataFusion weekly project plan (Andrew Lamb) - Feb 12, 2024

[alamb]

Follow on to Feb 5, 2024

It would be great for other contributors to DataFusion who plan non trivial work could try to make them visible somehow as well 🙏 (feel free to copy / modify the format)

My Highlights from Last Week

• Comet (the spark accelerator) completed IP clearance Initial PR datafusion-comet#1
• DataFusion SIGMOD paper about DataFusion was accepted
• We are getting close to beginning the large scale "extract scalar functions from the core" Organization of functions. Feedback most welcome
• Lots of good new background / docs about what a PruningPredicate is: (thanks @comphead, @Jefffrey and @appletreeisyellow for the reviews and feedback): example pruning.rs and docs

Boilerplate Introduction

The idea of this ticket is make my plans for DataFusion visible, largely for my own personal organizational needs, but also to:

1. Communicate / coordinate in the community
2. Help provide an interesting summary of what is happening in DataFusion this week

My (personal) goals for this project

1000 projects are built using DataFusion 📈

My (personal) plans for this week

• get one or two more PRs with more extracted functions [DISCUSS] organization of functions #9100 and file tickets to track the rest of the work
• Get PR ready for review to improve ClickBench performance Faster String Aggregation (this is my personal fun project and I am so close)
• More advanced pruning predicates, specifically Support "A column is known to be entirely NULL" in PruningPredicate #9171

Project Queue (list of future projects)

• SIGMOD paper feedback: Finalize SIGMOD 2024 paper ~(if accepted)~ #8373
• Organize agenda for DataFusion meetup UPDATE: VENUE IS FULL Invitation: Mar 25, 2024 DataFusion meetup before the DataCouncil March 2024 in Austin TX #8522
• Improved Statistics: Epic: Statistics improvements #8227 (specifically, Introduce a way to represent constrained statistics / bounds on values in Statistics #8078)

Projects I plan to help actively help review / plan

• Improved planning speed [Epic] A collection of issues to improve planning performance / speed / efficiency #5637 with @matthewmturner
• Hardening Recursive CTEs Add support for recursive CTEs #462 with @matthewgapp (All that remains is to get memory limiting in place and we could enable this by default). If I don't see motion this week I'll probably file a ticket to track the remaining work explicitly.

Algorithm for (my) prioritizing PR reviews

Note there are many committers who can and do review and merge PRs, so this is not the priorities of the project as a whole, just the approximate algorithm I use to prioritize my own time

Priority:

1. Bug fixes (where something is just incorrect), especially regressions (where it used to work and now does not)
2. Improvements directly related to features needed for InfluxDB (my employer)
3. Documentation and test improvements (I view these as very strategically important)
4. PRs that I think are strategically important
5. Other new features / additions to functionality (note this is the lowest)

The top strategically important projects in my head are:

• Anything that makes it easier to use DataFusion as a user (docs, examples, better APIs, etc), including Getting started guide for new users (who want to use DataFusion in their project)
• Anything that improves DataFusion's quality such as bug fixes, and improved / expanded tests, etc
• Making function packages modular and easier to mix/match: [EPIC] Unify Function Interface (remove BuiltInScalarFunction) #8045
• Improving performance of aggregation - especially high cardinality groupings, where we still lag state of the art by 2x

Thus, if you are interested in contributing to DataFusion and are interested in a fast turn around time I would recommend looking into bug fixes / test improvements / documentation or the projects named above

If you propose adding new functionality, the review cycle will likely be longer. You can make it a shorter cycle by looking at the comments on other recent PRs and following the same model (e.g. ensure there are tests in sqllogictest for example, the CI passes, includes documentation, etc)

[alamb]

DataFusion

• Add to_char function implementation using chrono formats #9181
• Add SessionContext::read_batches #9197
• Support compute return types from argument values (not just their DataTypes) #8985
• Add constant expression support to equivalence properties #9198
• feat: track memory consumers for GreedyMemoryPoolState #9015
• Add cargo audit CI #9182
• feat: implement select directly from s3 and gcs locations in datafusion-cli #9199
• Return an error when a column does not exist in window function #9202
• Add to_date function #9019

[alamb]

DataFusion

• chore(pruning): Support IS NOT NULL predicates in PruningPredicate #9208
• Partitioning fixes #9207
• feat: track memory consumers for GreedyMemoryPoolState #9015
• feat: Add support for Utf8Type and TimeStamp in Parquet statistics #9129
• feat: support FixedSizeList Type Coercion #9108
• Consolidate TreeNode transform and rewrite APIs #8891
• fix: issue #8838 discard extra sort when sorted element is wrapped #9127
• Add ForeignKey constraint type #8566
• Transform with payload #8664
• Fix incorrect results for NOT IN subqueries with nulls #8271

Watching

• Support compute return types from argument values (not just their DataTypes) #8985
• Consolidate TreeNode transform and rewrite APIs #8891
• Feat/make dfschema wrap schemaref #8905

Sqlparser

• Support DECLARE syntax for snowflake and bigquery datafusion-sqlparser-rs#1122

[alamb]

Review Queue

DataFusion

• fix: issue #9130 substitute redundant columns when doing cross join #9154 (review)
• fix: add ArrayAndElementAndOptionalIndex for proper casting in array_position #9233 (review)

Sqlparser:

• Support DECLARE syntax for snowflake and bigquery datafusion-sqlparser-rs#1122

[alamb]

BTW I am on vacation the week of Feb 19, so I will be reviewing stuff as I can. but likely will not make as much progress as most weeks.

I'll be back the following week as normal