The following should happens before:
- incorrect shards was selected
- "Cannot infer type of an empty tuple" exception for empty tuple() for localhost
- "Function tuple requires at least one argument" exception for empty tuple() for remote node
MSan reports [1]:
2021.04.17 15:20:06.665152 [ 56 ] {2336bf92-0269-4acd-8b3f-f09623223d18} <Debug> executeQuery: (from [::1]:44744, using production parser) SELECT * FROM dist_01757 WHERE dummy IN ('255', 0) FORMAT Null
...
0 0x305af885 in (anonymous namespace)::shardContains() obj-x86_64-linux-gnu/../src/Interpreters/OptimizeShardingKeyRewriteInVisitor.cpp:50:28
...
Uninitialized value was created by an allocation of 'sharding_value' in the stack frame of function '_ZN12_GLOBAL__N_113shardContainsERKN2DB5FieldERKNSt3__112basic_stringIcNS4_11char_traitsIcEENS4_9allocatorIcEEEERKNS4_10shared_ptrINS0_17ExpressionActionsEEESC_RKNS0_7Cluster9ShardInfoERKNS4_6vectorImNS8_ImEEEE'
0 0x305ae260 in (anonymous namespace)::shardContains() obj-x86_64-linux-gnu/../src/Interpreters/OptimizeShardingKeyRewriteInVisitor.cpp:42
[1]: https://clickhouse-test-reports.s3.yandex.net/23212/d02eb3f6b7a8dc07bbed5605e1afcf7634436b3a/fuzzer_msan/report.html#fail1
P.S. fuzzers are great!
v2: fix optimize_skip_unused_shards_rewrite_in for sharding_key wrapped into function
v3: fix column name for optimize_skip_unused_shards_rewrite_in
v4: fix optimize_skip_unused_shards_rewrite_in with Null
v5:
- squash with Remove query argument for IStreamFactory::createForShard()
- use proper column after function execution (using sharding_key_column_name)
- update the test reference since (X) now is tuple(X)
