add fuzzable arrow flight SQL server

[jayshrivastava]

This change adds an arrow flight sql server which can be targeted by SQLancer. It generates in-memory tables using DML.

These are the results

• Ran 3k queries
• 73% success rate (ie. 27% of queries ran into errors)
  • These can be ignored because sqlancer can produce invalid queries
• Importantly, did not generate any logs inlogs/datafusion_custom_log/error_report.log, meaning all validations passed

➜  target git:(main) java --add-opens=java.base/java.nio=org.apache.arrow.memory.core,ALL-UNNAMED -jar sqlancer-*.jar --random-seed 0 --num-threads 1 --max-generated-databases 1 --num-tries 10 --num-queries 500 datafusion
WARNING: Unknown module: org.apache.arrow.memory.core specified to --add-opens
WARNING: A terminally deprecated method in sun.misc.Unsafe has been called
WARNING: sun.misc.Unsafe::objectFieldOffset has been called by com.github.benmanes.caffeine.cache.UnsafeAccess (file:/Users/jayant.shrivastava/code/datafusion-sqllancer/target/lib/caffeine-2.9.3.jar)
WARNING: Please consider reporting this to the maintainers of class com.github.benmanes.caffeine.cache.UnsafeAccess
WARNING: sun.misc.Unsafe::objectFieldOffset will be removed in a future release
Nov 16, 2025 4:27:36 PM org.apache.arrow.driver.jdbc.shaded.org.apache.arrow.memory.BaseAllocator <clinit>
INFO: Debug mode disabled. Enable with the VM option -Darrow.memory.debug.allocator=true.
Nov 16, 2025 4:27:36 PM org.apache.arrow.driver.jdbc.shaded.org.apache.arrow.memory.DefaultAllocationManagerOption getDefaultAllocationManagerFactory
INFO: allocation manager type not specified, using netty as the default type
Nov 16, 2025 4:27:36 PM org.apache.arrow.driver.jdbc.shaded.org.apache.arrow.memory.CheckAllocator reportResult
INFO: Using DefaultAllocationManager at memory/netty/DefaultAllocationManagerFactory.class
[2025/11/16 16:27:41] Executed 296 queries (59 queries/s; 0.40/s dbs, successful statements: 66%). Threads shut down: 1.
[2025/11/16 16:27:46] Executed 804 queries (101 queries/s; 0.20/s dbs, successful statements: 70%). Threads shut down: 2.
[2025/11/16 16:27:51] Executed 1389 queries (117 queries/s; 0.40/s dbs, successful statements: 73%). Threads shut down: 4.
[2025/11/16 16:27:56] Executed 1969 queries (116 queries/s; 0.20/s dbs, successful statements: 74%). Threads shut down: 5.
[2025/11/16 16:28:01] Executed 2510 queries (108 queries/s; 0.40/s dbs, successful statements: 73%). Threads shut down: 7.
[2025/11/16 16:28:06] Executed 3109 queries (119 queries/s; 0.40/s dbs, successful statements: 73%). Threads shut down: 9.

➜  target git:(main) cat logs/datafusion_custom_log/error_report.log

Example DMLs

➜  target git:(main) cat logs/datafusion/database0-cur.log | grep DML
/*DML*/CREATE TABLE t0(v0 BIGINT, v1 STRING, v2 BOOLEAN, v3 DOUBLE);
/*DML*/INSERT INTO t0(v3, v2, v0) VALUES ('+Inf'::Double, true, -81839914);
/*DML*/INSERT INTO t0(v0) VALUES (-81839914);
/*DML*/INSERT INTO t0(v1, v0, v2) VALUES ('-81839914', -81839914, false);
/*DML*/INSERT INTO t0(v0, v1, v3) VALUES (-81839914, '', 0.9620689182075386), (412901507, '', '-Inf'::Double);
/*DML*/INSERT INTO t0(v0, v2) VALUES (43703267, false);
/*DML*/INSERT INTO t0(v2) VALUES (false);
/*DML*/INSERT INTO t0(v2, v3, v1) VALUES (false, 0.3227677982432213, '2v');
/*DML*/INSERT INTO t0(v3) VALUES (0.6808360271610692);
/*DML*/INSERT INTO t0(v3) VALUES ('NaN'::Double);
/*DML*/INSERT INTO t0(v0, v2) VALUES (3, true), (-4, true);
/*DML*/CREATE TABLE t0_stringview AS SELECT v0, arrow_cast(v1, 'Utf8View') as v1, v2, v3 FROM t0;

Example queries

➜  target git:(main) tail logs/datafusion/database0-cur.log
select * from tt;
SELECT ALL LAST_VALUE(SQRT(tt0.v0)) OVER (PARTITION BY tt0.v2, tt0.v1) FROM t0_stringview AS tt0 WHERE tt0.v2;
SELECT tt1.v3, tt0.v3, ARROW_CAST(tt1.v1, 'Utf8'), tt1.v0, ARROW_CAST(tt0.v1, 'Utf8'), tt0.v2 FROM t0_stringview AS tt0 FULL OUTER JOIN t0_stringview AS tt1 ON false WHERE ((TRANSLATE(REGEXP_MATCH(tt0.v1, tt0.v1, tt1.v0), tt1.v1, ((tt1.v1)~*(tt0.v1))))<(((REGEXP_REPLACE(tt1.v1, tt0.v1))!~(((tt0.v1)!~~(tt1.v1))))));
SELECT COUNT(*) FROM t0_stringview AS tt0 WHERE ((false)OR(STARTS_WITH((('!喐') ILIKE ('24\n')), LTRIM(1))));
SELECT tt0.v0, tt1.v3, tt0.v2, tt0.v3, ARROW_CAST(tt1.v1, 'Utf8'), tt1.v2 FROM t0_stringview AS tt0 FULL OUTER JOIN t0 AS tt1 ON tt0.v2 GROUP BY FIND_IN_SET('Zm', tt1.v3) ORDER BY tt0.v0 ASC;
select * from tt;
SELECT ALL 1.7323877647274428E308 OVER () FROM t0_stringview AS tt0 WHERE ((0.48131556370595674)IS DISTINCT FROM(CEIL(ABS(tt0.v2))));
SELECT DISTINCT ((((-4)|(-5)))<<((- -2.091572074E9))) FROM t0_stringview AS tt0 WHERE false GROUP BY '2046108624', -1;
SELECT COUNT(*) FROM t0_stringview AS tt0 INNER JOIN t0 AS tt1 ON ((SUBSTR(tt1.v3, tt1.v0))=(tt0.v1)) WHERE ((((tt1.v1)!~~(RTRIM(tt0.v1))))<(((tt0.v1)||(tt0.v1))));
SELECT DISTINCT tt1.v3, tt1.v0 FROM t0 AS tt0 CROSS JOIN t0 AS tt1 GROUP BY false, (NOT (((('oV')>('-81839914')))IS DISTINCT FROM(false)));

Example log file:
database0-cur.log