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

[alamb]

Follow on to #9899

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

• 37.0.0 was released Release DataFusion 37.0.0 #9682 🚀
• DataFusion WASM is becoming a thing @milenkovicm showed us how to make WASM UDFs WASM UDFs #9326 / https://github.com/milenkovicm/wasaffi and @waynxia showed off a WASM playground: Web Playground with WASM #9834
• @peter-toth made LogicalPlan rewriting with TreeNode easier and faster: Consistent LogicalPlan subquery handling in TreeNode::apply and TreeNode::visit  #9913
• @jayzhan211 began pulling aggregate functions out of the core: Move First Value UDAF and builtin first / last function to aggregate-functions #9960
• "Extract scalar functions from the core" continues: [Epic] Port BuiltInFunctons to datafusion-functions-* crates #9285 (thanks @erenavsarogullari and @Omega359 for keeping the 🚋 running here)
• Started talking about indexing: Indexing Support in DataFusion? #9963
• Started talking about ListingTable: [EPIC] Improve the performance of ListingTable #9964

My (personal) goals

1000 projects are built using DataFusion 📈

My (personal) plans for this week

• Release version 0.45.0 datafusion-sqlparser-rs#1162
• Reviews !
• Make query planning (at least) 2x faster [EPIC] Stop copying LogicalPlan during OptimizerPasses #9637)
• Document ParquetExec more (so it is clear / easier to understand what is already supported and what else could be useful API in ParquetExec to pass in RowSelections to ParquetExec (enable custom indexes, finer grained pushdown) #9929)

Project Queue (list of future projects)

• API in ParquetExec to pass in RowSelections to ParquetExec (enable custom indexes, finer grained pushdown) #9929

Projects I plan to help actively help review / plan

• [EPIC] Implement StringViewArray and BinaryViewArray arrow-rs#5374
• [Epic] Port BuiltInFunctons to datafusion-functions-* crates #9285
• [EPIC] Support converting Exprs and LogicalPlans --> SQL Strings #9494
• Supporting using parallel parquet writer outside of Datafusion query execution #9493 with @wiedld

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

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)

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]

@Omega359 get the 🍀 with lucky PR 10,000: #10000

[Omega359]

It wasn't on purpose :)

[alamb]

Review Queue:

DataFusion

• move the Log, Power functions to datafusion-functions #9983
• Further clarification of the supports_filters_pushdown documentation #9988
• [RFC] Register scalars with boxed fn impl #9980
• move Trunc, Cot, Round, iszero functions to datafusion-functions #10000
• move the Log, Power functions to datafusion-functions #9983
• Prune columns are all null in ParquetExec by row_counts , handle IS NOT NULL #9989

arrow-rs/parquet:

• Add tests for Arrow Flight support for StringViewArray and BinaryViewArray arrow-rs#5601
• feat: support reading and writingStringView and BinaryView in parquet (part 2) arrow-rs#5557
• feat: implemented with_field() for FixedSizeListBuilder arrow-rs#5541

[alamb]

BTW I don't think I say it enough, but @comphead I really appreciate all the effort you put in to review PRs

[alamb]

Review Queue:

DataFusion

• Introduce TreeNode::exists() API, avoid copying expressions #10008
• fix: reduce lock contention in RepartitionExec::execute #10009

sqlparser-rs:

• support clickhouse parameterized views datafusion-sqlparser-rs#1181
• Support MySQL UNIQUE table constraint datafusion-sqlparser-rs#1164
• feat: support tailing commas on snowflake dialect. datafusion-sqlparser-rs#1205
• Fix bug to allow parenthesis query after INSERT INTO datafusion-sqlparser-rs#1157
• Support CREATE/DROP SECRET for duckdb dialect datafusion-sqlparser-rs#1208
• BigQuery: support unquoted hyphen in table/view declaration datafusion-sqlparser-rs#1178
• BigQuery: support unquoted hyphen in table/view declaration datafusion-sqlparser-rs#1178

arrow-rs/parquet:

• feat: support reading and writingStringView and BinaryView in parquet (part 2) arrow-rs#5557

[alamb]

Review Queue:

DataFusion

• Fix datafusion-cli cursor isn't on the right position in windows 7 cmd #10028 (review)
• Improve the performance of ltrim/rtrim/btrim #10006
• Add support_ignore_nulls and support_ordering in Aggregate expression #9991
• fix: RepartitionExec metrics #10025
• modify emit() of TopK to emit on batch_size rather than batch_size-1 #10030
• feat: eliminate redundant sorts on monotonic expressions #9813

arrow-rs/parquet:

• feat: encode FixedSizeBinary in JSON as hex string arrow-rs#5622 (review)

[alamb]

Review Queue:

DataFusion

• Fix DistinctCount for timestamps with time zone #10043
• Refactor physical create_initial_plan to iteratively & concurrently construct plan from the bottom up #10023
• fix: reduce lock contention in distributor channels #10026
• feat: optimize lower and upper functions #9971

arrow-rs/parquet:

• Support non-contiguous put payloads / vectored writes (#5514) arrow-rs#5538

[alamb]

Review Queue:

DataFusion

• Prune pages are all null in ParquetExec by row_counts and fix NOT NULL prune #10051
• Minor: get mutable ref to SessionConfig in SessionState #10050
• feat: support unnest multiple arrays #10044
• fix: reduce lock contention in distributor channels #10026
• feat: optimize lower and upper functions #9971
• fix: Support Dict types in in_list physical plans #10031

arrow-rs/parquet:

• feat: support reading and writingStringView and BinaryView in parquet (part 2) arrow-rs#5557
• Implement filter kernel for byte view arrays. arrow-rs#5624
• feat(ffi): add run end encoded arrays arrow-rs#5632

sqlparser-rs:

• Enhancing Trailing Comma Option datafusion-sqlparser-rs#1212
• Add support for quoted string backslash escaping datafusion-sqlparser-rs#1177
• support clickhouse parameterized views datafusion-sqlparser-rs#1181
• Fix dollar quoted string tokenizer datafusion-sqlparser-rs#1193

[alamb]

4 BuiltInScalarFunctions left

[liukun4515]

4 BuiltInScalarFunctions left

Hi @alamb
I scan some PRs about #9285, and find the the item of datafusion/proto/proto/datafusion.proto was deleted.

I have the question about this, how to serialize and deserialize the other language logical plan/expr to datafusion?

[alamb]

I have the question about this, how to serialize and deserialize the other language logical plan/expr to datafusion?

I @liukun4515 -- the logical plan serialization / deserialization handles ScalarUDF and other user defined functions as before.

So for example, here is an example of how functions are serialized:

https://github.com/apache/arrow-datafusion/blob/92318c808666695fb5c2f358083f1cb13158dca9/datafusion/proto/tests/cases/roundtrip_logical_plan.rs#L1867-L1892

[alamb]

Next week in #10172