V0.0.4

Welcome to Kùzu 0.0.4! In this release, we've introduced following improvements and features:

• Improved the performance of ingesting both node and rel tables for quite a bit.
• Supported undirected relationships in queries.
• Supported shortest path queries, and also improved variable-length joins.
• Added new data types, including SERIAL and STRUCT.
• Made Kùzu available on windows platforms.
• Added C and nodejs client APIs.

Please see our blog post for more details. Enjoy your reading!

What's Changed

• Update README.md by @andyfengHKU in #1454
• Fix set number threads by @acquamarin in #1456
• Fix fedora compiling by @ray6080 in #1458
• Add struct data type front end by @acquamarin in #1457
• fixing issue 1404 by @semihsalihoglu-uw in #1460
• Upgrade CI runner version for benchmark docker by @mewim in #1463
• Specify column types in csv reader and remove arrow status return value from functions by @ray6080 in #1461
• Fix compilation on fedora 38 by @ray6080 in #1467
• Rework hashjoin by @andyfengHKU in #1465
• Rework arrow array copy in node table copier by @ray6080 in #1469
• Fix issue 1466 by @ray6080 in #1470
• Shortest path planning by @andyfengHKU in #1472
• Shortest Path changes by @anuchak in #1468
• Fix macOS compilation issue by @mewim in #1473
• Add C API bindings by @mewim in #1471
• Rework node copy task by @ray6080 in #1475
• Add struct literal support by @acquamarin in #1462
• Update README.md python code example by @wenhoujx in #1478
• Skip first pass read of parquet data by @ray6080 in #1480
• Struct planning by @andyfengHKU in #1479
• Fix incorrect cast int16 func name by @ray6080 in #1484
• Add c, cpp examples by @andyfengHKU in #1485
• Add int16 int32 support to get_as_arrow. by @wenhoujx in #1483
• Remove PropertyNameDataType by @ray6080 in #1488
• Fix process fork support for Python API by @mewim in #1487
• Encapsulate templatizing of HashIndexBuilder inside PrimaryKeyIndexBuilder by @ray6080 in #1491
• Add loader and storage support for struct by @acquamarin in #1490
• SSP pipeline by @andyfengHKU in #1489
• Add SERIAL as node table primary key by @ray6080 in #1493
• Add sink before recursive join by @andyfengHKU in #1500
• Update cost model for shortest path by @andyfengHKU in #1499
• Node.js API prototyping by @ynoza in #1501
• Add sideway information passing for shortest path extend by @andyfengHKU in #1502
• Vlq rework by @andyfengHKU in #1505
• List valueVector refactor by @acquamarin in #1503
• Issue 951 by @andyfengHKU in #1508
• Fix bug in var length queries sorting by @anuchak in #1509
• Fix struct pack bug by @acquamarin in #1515
• Add aggreate key dependency optimizer by @andyfengHKU in #1517
• Auto-parse end2end tests by @rfdavid in #1507
• Add support to struct of list by @acquamarin in #1518
• Regex String Functions by @gaurav8297 in #1516
• Remove large rel list by @ray6080 in #1520
• Convert e2_read_list test to the new framework by @rfdavid in #1523
• Support null values in list by @acquamarin in #1524
• Add document for PyG remote backend by @mewim in #1525
• Add STRUCT binding for Python API by @mewim in #1527
• Converted order_by and copy tests to the new format by @rfdavid in #1526
• Add struct binding for C API by @mewim in #1531
• Add null to struct fields by @acquamarin in #1533
• Allow passing a test as an argument to e2e_read_test by @rfdavid in #1535
• Add list_sum and list_sort functions by @gaurav8297 in #1529
• Undirected query logical changes by @aziz-mu in #1528
• Rework column scan and lookup functions by @ray6080 in #1539
• List field name to idx cache by @acquamarin in #1534
• Rework column write function by @ray6080 in #1540
• Recursive path by @andyfengHKU in #1538
• More list functions by @gaurav8297 in #1543
• Undirected query for single label Nodes/Rels by @aziz-mu in #1542
• Move scans of tableID into node/relIDs out of Column by @ray6080 in #1550
• Separate null bits from data in columns by @ray6080 in #1545
• Recursive join no path tracking optimizer by @andyfengHKU in #1547
• Physical type by @acquamarin in #1549
• Undirected query for multi-labelled and unlabelled nodes. by @aziz-mu in #1552
• Add recursive rel logical type by @andyfengHKU in #1553
• Fix issue 1546 by @andyfengHKU in #1554
• Add visitor pattern to bound statement by @andyfengHKU in #1557
• Push recursive path length into operator by @andyfengHKU in #1555
• Testing framework v2 by @rfdavid in #1548
• Refactor commit/rollback/checkpoint interface by @ray6080 in #1558
• Refactor extend direction & rel data direction by @andyfengHKU in #1560
• Rework ResultSetDescriptor by @andyfengHKU in #1561
• Change hash index concurrent build by @ray6080 in #1564
• Check statement read only through visitor pattern by @andyfengHKU in #1559
• Enable multiple test cases inside one test file by @rfdavid in #1567
• add LSQB queries as test by @aziz-mu in #1566
• Add loader support for struct of list/struct/string by @acquamarin in #1563
• Use cross-platform glob library by @benjaminwinger in #1562
• Convert binder error tests to end to end tests by @rfdavid in #1569
• Fix parser error by @acquamarin in #1572
• Add Windows Support by @benjaminwinger in #1573
• Fix win ci by @andyfengHKU in #1576
• Fix issue-1575 by @andyfengHKU in #1579
• Convert demo db tests to e2e tests by @rfdavid in #1574
• Track edge ID in recursive join by @andyfengHKU in #1565
• Convert lsqb cpp tests to end to end tests by @rfdavid in #1580
• Fix return functor windows compatibility by @andyfengHKU in #1586
• NPY loader fix by @acquamarin in #1571
• Decrease Windows CI runner thread by @mewim in #1591
• Convert e2e_exception_test to end to end test by @rfdavid in #1584
• Fix issue-1512 by @andyfengHKU in #1593
• Update Slack links by @mewim in #1594
• Add LDBC interactive short queries to test by @aziz-mu in #1589
• Convert e2e_update_node.cpp to e2e test files by @rfdavid in #1597
• Multi label recursive join by @andyfengHKU in #1581
• Don't always set GENERIC_READ when opening files on windows by @benjaminwinger in #1585
• Use GetFileSizeEX instead of GetFileSize on windows by @benjaminwinger in #1587
• Rename e2e_read_test to e2e_test by @rfdavid in #1599
• Print error details on test failure by @rfdavid in #1598
• Allow writing test queries in multiple lines by @rfdavid in #1602
• Add more test cases from documentation by @andyfengHKU in #1604
• Support more native parquet types by @acquamarin in #1600
• Add option -SKIP to skip test cases by @rfdavid in #1603
• Fix compilation issue on mac by @acquamarin in #1606
• Add c api documentation by @acquamarin in #1608
• Recursive join fix by @andyfengHKU in #1607
• Build Python wheels on Windows by @benjaminwinger in #1595
• Node js binding rework by @mewim in #1613
• Make Node.js build optional by @mewim in #1615
• value refactor by @acquamarin in #1614
• add update to table with serial by @acquamarin in #1610
• Fix struct parsing by @andyfengHKU in #1609
• Port bundled linenoise to windows by @benjaminwinger in #1612
• Fix spdlog header resolution by @mewim in #1620
• Add CI build scripts for Node.js by @mewim in #1622
• Auto-select precompiled bin for Node.js install by @mewim in #1625
• Serial query processing by @acquamarin in #1624
• Add documentations for Node.js API by @mewim in #1626
• Add assertion import for Node.js API by @mewim in #1629
• Rework rel copy task scheduling and improve copy performance by @ray6080 in #1621
• Add constant to test_files directory & fix directory by @rfdavid in #1631
• update readme by @andyfengHKU in #1623

New Contributors

• @wenhoujx made their first contribution in #1478
• @gaurav8297 made their first contribution in #1516
• @benjaminwinger made their first contribution in #1562

Full Changelog: v0.0.3...v0.0.4