DataFusion weekly project plan (Andrew Lamb) - July 8, 2024

[alamb]

Follow on to #11190

My (personal) North ⭐ : 1000 projects are built using DataFusion 📈

It would be great for other contributors to DataFusion to to make them visible somehow as well 🙏 (feel free to copy / modify the format of this ticket)

Externally Facing News

DataFusion mentioned in "What goes around, comes around... and around...." paper from Andy Pavlo and Mike Stonebreaker
https://twitter.com/andy_pavlo/status/1807799839616614856

Feature Highlights

• Introduce user defined SQL planner API from @jayzhan211
• docs: add example for custom file format with COPY TO @tshauk
• @lewiszlw continues to make the code better Enable clone_on_ref_ptr clippy lint on execution crate #11239 ❤
• @Blizzara and @Lordworms continue to bash out substrait support (they are up to tpch q4 and 5)! Implement TPCH substrait integration teset, support tpch_3 #11298
• @jayzhan211 supported DUCKDB style struct syntax: Support DuckDB style struct syntax #11214

> select {'foo': 1};
+------------------------------------+
| named_struct(Utf8("foo"),Int64(1)) |
+------------------------------------+
| {foo: 1}                           |
+------------------------------------+
1 row(s) fetched.
Elapsed 0.015 seconds.

> select {'foo': {'bar': 2}};
+--------------------------------------------------------------+
| named_struct(Utf8("foo"),named_struct(Utf8("bar"),Int64(2))) |
+--------------------------------------------------------------+
| {foo: {bar: 2}}                                              |
+--------------------------------------------------------------+
1 row(s) fetched.
Elapsed 0.002 seconds.

My (personal) plans for this week

• SQLParser release: Release sqlparser-rs version 0.48.0 datafusion-sqlparser-rs#1296
• "move the needle" projects: improve docs + writing [EPIC] Improve examples to make them easier to navigate #11172 Blog post with DataFusion Jan - June 2024 #9602,
• (Fun Project): Faster multi-column aggregation with @jayzhan211 Improve performance for grouping by variable length columns (strings) #9403 (Optionally display schema in explain plan #11177 was a step towards this)
• July 10, 2024 ASF Board Report #10282

Project Queue (list of future projects)

Projects I plan to help actively help review / plan

• Complete SQL planner extension API Discussion: make it easier for specify SQL --> function translation #10534 with @jayzhan211 and @samuelcolvin, @xinlifoobar, @dharanad)

• Various upstream parquet metadata changes: Add ParquetMetadataWriter allow ad-hoc encoding of ParquetMetadata arrow-rs#6000, Do not write ColumnIndex for null columns when not writing page statistics arrow-rs#6011, Add size statistics to ParquetMetaData introduced in PARQUET-2261 arrow-rs#5486,

• Help push user defined aggregates over the line ([Epic] Unify AggregateFunction Interface (remove built in list of AggregateFunction s), improve the system #8708) - thanks to @Rachelint and @jcsherin)

• [EPIC] Continued correct and improved extracting Parquet statistics into ArrayRefs  #10922 (kudos to @Rachelint and @efredine) most of what is left is porting it upstream to arrow-rs)

• [EPIC] Implement StringViewArray and BinaryViewArray arrow-rs#5374 and [Epic] Implement support for StringView in DataFusion #10918 with @XiangpengHao and others

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
• Improved performance (with benchmarks)

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, especially if the PR is large/complex and not connected to a wider need, 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)

Background

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

[alamb]

Review Queue

Arrow

• Add ParquetMetadataWriter allow ad-hoc encoding of ParquetMetadata arrow-rs#6000
• Fix some unsound unsafe code in arrow and attempt to improve safety docs. arrow-rs#6021
• Do not write ColumnIndex for null columns when not writing page statistics arrow-rs#6011
• Add begin/end_transaction methods in FlightSqlServiceClient arrow-rs#6026
• Add size statistics to ParquetMetaData introduced in PARQUET-2261 arrow-rs#5486
• Make display of interval types more pretty arrow-rs#6006

DataFusion

• fix: Fix eq properties regression from #10434 #11363
• Improve CommonSubexprEliminate rule with surely and conditionally evaluated stats #11357
• Implement prettier SQL unparsing (more human readable) #11186
• Coerce types for all union children plans when eliminating nesting #11386
• Implement ScalarFunction MAKE_MAP and MAP #11361
• Change array_agg to return null on no input rather than empty list #11299
• Add customizable equality and hash functions to UDFs #11392
• Extract parquet statistics for StructArray #11289
• feat: add raw aggregate udf planner #11371
• Introduce the DynamicFileCatalog in datafusion-catalog #11035
• Feat: Implement hf:// / "hugging face" integration in datafusion-cli #10792
• Support SortMergeJoin spilling #11218

[alamb]

I am so behind on reviews. I hope to try and catch up tomorrow

[alamb]

Review Queue

Arrow

• Add parquet StatisticsConverter for arrow reader arrow-rs#6046
• Update Parquet thrift generated structures arrow-rs#6045
• Directly decode String/BinaryView types from arrow-row format arrow-rs#6044
• Remove impl<T: AsRef<[u8]>> From<T> for Buffer that easily accidentally copies data arrow-rs#6043
• Unsafe improvements: arrow/{array_reader,buffer}. arrow-rs#6025
• Fix some unsound unsafe code in arrow and attempt to improve safety docs. arrow-rs#6021

DataFusion

• Add SessionStateBuilder and extract out the registration of defaults #11403
• Add ArrowBytesViewMap and ArrowBytesViewSet #11421
• Support SortMergeJoin spilling #11218
• Extract parquet statistics for StructArray #11289
• feat: add raw aggregate udf planner #11371
• Add extension hooks for encoding and decoding UDAFs and UDWFs #11417
• Introduce the DynamicFileCatalog in datafusion-catalog #11035
• Feat: Implement hf:// / "hugging face" integration in datafusion-cli #10792
• Support SortMergeJoin spilling #11218

[alamb]

Review Queue

Arrow

• Implement min max support for string/binary view types arrow-rs#6053
• Add parquet StatisticsConverter for arrow reader arrow-rs#6046
• Fix some unsound unsafe code in arrow and attempt to improve safety docs. arrow-rs#6021

DataFusion:

• Plan LATERAL subqueries #11456
• Add ArrowBytesViewMap and ArrowBytesViewSet #11421
• Extract parquet statistics for StructArray #11289
• Introduce the DynamicFileCatalog in datafusion-catalog #11035
• Feat: Implement hf:// / "hugging face" integration in datafusion-cli #10792

[alamb]

Next week #11474