[ARROW-3338] [Python] Crash when schema and columns do not match

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.11.0
• Component/s: C++, Python
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/19671
• Assignee: Krisztian Szucs
• Reporter: Joseph Toth
• Created: 27/Sep/18 03:48
• Updated: 11/Jan/23 07:26
• Resolved: 27/Sep/18 14:58

Description

I get a segfault when the arrays do not match the schema. Running latest from HEAD ee9b1ba426e2f1f117cde8d8f4ba6fbe3be5674c

sch = pa.schema([
  pa.field('a', pa.int64()),
  pa.field('b', pa.string()),
])
rb = pa.RecordBatch.from_arrays([pa.array([1])], ['a'])
t = pa.Table.from_batches([rb], sch)

In[6]: t = pa.Table.from_batches([rb], sch)
[mine:11188] *** Process received signal ***
[mine:11188] Signal: Segmentation fault (11)
[mine:11188] Signal code: Address not mapped (1)
[mine:11188] Failing at address: 0x49
[mine:11188] [ 0] /lib/x86_64-linux-gnu/libpthread.so.0(+0x128e0)[0x7f89d7a578e0]
[mine:11188] [ 1] /usr/local/lib/python3.6/dist-packages/pyarrow/libarrow.so.10(_ZNK5arrow17SimpleRecordBatch6columnEi+0x48)[0x7f896d651548]
[mine:11188] [ 2] /usr/local/lib/python3.6/dist-packages/pyarrow/libarrow.so.10(_ZN5arrow5Table17FromRecordBatchesERKSt10shared_ptrINS_6SchemaEERKSt6vectorIS1_INS_11RecordBatchEESaIS8_EEPS1_IS0_E+0x48a)[0x7f896d65771a]
[mine:11188] [ 3] /usr/local/lib/python3.6/dist-packages/pyarrow/lib.cpython-36m-x86_64-linux-gnu.so(+0x13e98c)[0x7f896dd8a98c]
[mine:11188] [ 4] /usr/bin/python3.6[0x50c3e5]
[mine:11188] [ 5] /usr/bin/python3.6(_PyEval_EvalFrameDefault+0x449)[0x50fad9]
[mine:11188] [ 6] /usr/bin/python3.6[0x50dee7]
[mine:11188] [ 7] /usr/bin/python3.6[0x51b2ea]

Issue Links

• links to: GitHub Pull Request #2643