SEGV in reading parquet data if column type mismatch occurs

[czentgr]

Bug description

The issue is easily reproducible:

make a table that contains a single varchar column

create table test (c0 varchar) with (format='PARQUET');
insert into test values ('abc');

Save the resulting parquet file and delete the table.
Create a new table:

create table test (c0 integer) with (format='PARQUET');

and copy the data file into the storage path of the new table.

This simulates when data is created externally and a new table is created to process it.
If the table definition is incorrect then a SEGV can occur.

A basic read query:

select * from test;

causes the following SEGV:

I20250214 11:04:53.720890 67128642 TaskManager.cpp:588] Starting task 20250214_032653_00006_jukim.1.0.0.0 with 16 max drivers.
*** Aborted at 1739549093 (unix time) try "date -d @1739549093" if you are using GNU date ***
PC: @                0x0 facebook::velox::tpch::dbgen::name_bits
*** SIGSEGV (@0xc) received by PID 44213 (TID 0x16ebd7000) stack trace: ***
    @        0x187e06e04 _sigtramp
    @        0x10c929e34 facebook::velox::ByteOutputStream::appendOne<>()
    @        0x10c929e34 facebook::velox::ByteOutputStream::appendOne<>()
    @        0x10c911150 facebook::velox::serializer::presto::detail::VectorStream::appendLength()
    @        0x10c926f2c facebook::velox::serializer::presto::detail::VectorStream::appendLengths<>()
I20250214 11:04:53.750250 67134468 TaskManager.cpp:634] No more splits for 20250214_032653_00006_jukim.1.0.0.0 for node 0
    @        0x10c926bd8 facebook::velox::serializer::presto::detail::(anonymous namespace)::appendStrings()
    @        0x10c9206fc facebook::velox::serializer::presto::detail::(anonymous namespace)::serializeFlatVector<>()
    @        0x10c91e744 _ZZZN8facebook5velox10serializer6presto6detail15serializeColumnERKNSt3__110shared_ptrINS0_10BaseVectorEEERKN5folly5RangeIPKiEEPNS3_12VectorStreamERNS0_7ScratchEENK3$_0clEvENKUlvE_clEv
    @        0x10c90f564 facebook::velox::serializer::presto::detail::serializeColumn()::$_0::operator()()
    @        0x10c90f3ac facebook::velox::serializer::presto::detail::serializeColumn()
    @        0x10c910054 facebook::velox::serializer::presto::detail::(anonymous namespace)::serializeWrapped()
    @        0x10c90f468 facebook::velox::serializer::presto::detail::serializeColumn()
    @        0x10c8e1efc facebook::velox::serializer::presto::detail::PrestoIterativeVectorSerializer::append()
    @        0x10d766404 facebook::velox::VectorStreamGroup::append()
    @        0x10cc61614 facebook::velox::exec::detail::Destination::advance()
    @        0x10cc64424 facebook::velox::exec::PartitionedOutput::getOutput()
    @        0x10ca631f0 facebook::velox::exec::Driver::runInternal()::$_9::operator()()
    @        0x10ca50498 facebook::velox::exec::Driver::withDeltaCpuWallTimer<>()
    @        0x10ca4e598 facebook::velox::exec::Driver::runInternal()
    @        0x10ca507a4 facebook::velox::exec::Driver::run()
    @        0x10ca53f04 facebook::velox::exec::Driver::enqueue()::$_0::operator()()
    @        0x10ca53e14 folly::detail::function::call_<>()
    @        0x102afa06c folly::detail::function::FunctionTraits<>::operator()()
    @        0x10db250b4 folly::ThreadPoolExecutor::runTask()
    @        0x10da92fe0 folly::CPUThreadPoolExecutor::threadRun()
    @        0x10db2b18c _ZNSt3__18__invokeB8ne180100IRMN5folly18ThreadPoolExecutorEFvNS_10shared_ptrINS2_6ThreadEEEERPS2_JRS5_EvEEDTcldsdeclsr3stdE7declvalIT0_EEclsr3stdE7declvalIT_EEspclsr3stdE7declvalIT1_EEEEOSD_OSC_DpOSE_
    @        0x10db2b0dc _ZNSt3__115__apply_functorB8ne180100IMN5folly18ThreadPoolExecutorEFvNS_10shared_ptrINS2_6ThreadEEEENS_5tupleIJPS2_S5_EEEJLm0ELm1EENS8_IJEEEEENS_13__bind_returnIT_T0_T2_Xsr22__is_valid_bind_returnISD_SE_SF_EE5valueEE4typeERSD_RSE_NS_15__tuple_indicesIJXspT1_EEEEOSF_
    @        0x10db2b07c _ZNSt3__16__bindIMN5folly18ThreadPoolExecutorEFvNS_10shared_ptrINS2_6ThreadEEEEJPS2_RS5_EEclB8ne180100IJEEENS_13__bind_returnIS7_NS_5tupleIJS8_S5_EEENSD_IJDpOT_EEEXsr22__is_valid_bind_returnIS7_SE_SI_EE5valueEE4typeESH_
    @        0x10db2ae5c folly::detail::function::call_<>()
    @        0x102afa06c folly::detail::function::FunctionTraits<>::operator()()
    @        0x102afa038 _ZZN5folly18NamedThreadFactory9newThreadEONS_8FunctionIFvvEEEENUlvE_clEv
    @        0x102af9fcc _ZNSt3__18__invokeB8ne180100IZN5folly18NamedThreadFactory9newThreadEONS1_8FunctionIFvvEEEEUlvE_JEEEDTclclsr3stdE7declvalIT_EEspclsr3stdE7declvalIT0_EEEEOS8_DpOS9_

An error might not always occur, for example

select max(c0) from test;

will succeed.

System information

n/a

Relevant logs