May 2025 Update

This update was generated with the assistance of AI. While we strive for accuracy, please note that AI-generated content may not always be error-free. We encourage you to verify any information that is important to you.

Core Library

• Add task barrier. #13087

• Add task barrier support for streaming aggregation, unnest and index join. #13273, #13293, #13244

• Add support for memory pool priority. #13386

• Support buffering in local exchange operator. #13234

• Add PageSpill for OutputBuffer spill. #13305

• Add lazy start with spill for LocalMerge. #13337

• Add support for leftSemiProject join in nested loop join. #12172

• Fix crash when aggregate push down applied on updated column with sparse row set. #13503

• Fix the timing order of CpuWallTimer. #13313

• Fix overflow in NegatedBigintValuesUsingHashTable::testInt64Range. #13523

Presto Functions

• Add simplify_geometry(), geometry_invalid_reason(), ST_IsValid(), ST_IsSimple(), ST_Point(), ST_X(), ST_Y(), ST_area() functions.

• Add bing_tile_at(), bing_tiles_around() functions.

• Add qdigest_agg(), cosine_similarity() functions.

• Add noisy_count_if_gaussian() function.

• Add quantile_at_value(), trimmed_mean() functions.

• Add geometry functions for WKT/WKB conversion.

• Add xxhash64_internal() with extended type support.

• Fix undefined behavior in qdigest when total weight exceeds int64_t max. #13336

• Fix handing of empty arrays in array_min(), array_max_by() functions. #13272

• Fix overflow check in timestamp addition. #13444

• Fix overflow in from_unixtime(). #13262

Spark Functions

• Add sqrt(), luhn_check() functions.

• Add CAST(bool as timestamp) support.

• Add support for legacy behavior in covariance functions. #12994

• Fix casting complex types to only cast recursively if child type changes. #13245

• Fix unescape json elements in array_join(). #13222

• Fix the duplicate map key handling for map() function. #13183

Connectors

• Support bucket write with non-partitioned table. #13283

• Support delta update on bucket column. #13404

• Support Null Column Projection in Batch Reader Adapter. #13430

• Add ColumnReaderOptions for better reader configuration. #12840

• Add support for float-to-double schema evolution. #13317

• Fix selective flatmap column reader read offset when all rows filtered out. #13350

• Fix access after buffer boundary causing crash in selective reader. #13344

• Fix crash if table column type does not match file column type. #12350

• Fix incorrect filter result during schema evolution when range is outside of old type. #13459

Performance and Correctness

• Optimize selective ARRAY and MAP reader. #13240

• Avoid decompressing data when estimating row size. #13365

• Clear hash table as soon as probe finish. #13254

• Allow empty file in table scan. #13241

Credits

aditi-pandit, ajeyabsf, Ali LeClerc, alileclerc, Amit Dutta, Anders Dellien, Andrii Rosa, arnavb, Artem Selishchev, Bikramjeet Vig, Chandrashekhar Kumar Singh, Chengcheng Jin, Christian Zentgraf, Deepak Majeti, Devavret Makkar, duanmeng, Eric Jia, Haiping Xue, Heidi Han, Jacob Khaliqi, Jacob Wujciak-Jens, James Gill, Jialiang Tan, Jimmy Lu, Ke Jia, Ke Wang, Kent Yao, Kevin Wilfong, Kien Nguyen, Kk Pulla, Kostas Xirogiannopoulos, Krishna Pai, lingbin, Lukas Krenz, MacVincent Agha-Oko, Mario Ruiz, Mingyu Zhang, Natasha Sehgal, Nathan Phan, NEUpanning, Oliver Xu, Patrick Sullivan, Pedro Eugenio Rocha Pedreira, Peter Enescu, Pradeep Vaka, Qian Sun, Rui Mo, Serge Druzkin, Shakyan Kushwaha, Soumya Duriseti, Surbhi Vijayvargeeya, Tanay Bhartia, Wei He, Xiao Du, Xiaoxuan Meng, Yabin Ma, Yenda Li, Yi Cheng Lee, Zhenyuan Zhao, Zhiguo Wu, Zhiying Liang