Crash with access violation (exit code -1073741819 (0xC0000005)) on Windows 10

[Mereep]

# Your code here
import pandas as pd
pd.__version__
Out[3]: '0.24.2'
t1 = {'a': 1, 'b': 1}
t2 = {'a': 2, 'b': 2}
test = [t1, t2],  # note the errernous comma at the end of the line 
type(test)
Out[7]: tuple
test
Out[8]: ([{'a': 1, 'b': 1}, {'a': 2, 'b': 2}],)
pd.DataFrame(test)
Process finished with exit code -1073741819 (0xC0000005)

Short:

pd.DataFrame(tuple([[1],[2]]))
Process finished with exit code -1073741819 (0xC0000005)

Problem description

Problem was caused in my case by having an additional comma at the end of a line that caused
a list of dicts being a tuple of list of dicts which caused that behaviour.

Althought the argument should not be accepted by Pandas, Pandas should raise an appropriate Exception instead of tearing Python down.

[jbrockmendel]

Can you try this on a newer version of pandas?

[Mereep]

New version seems to fix that

t1 = {'a': 1, 'b': 1}
t2 = {'a': 2, 'b': 2}
test = [t1, t2],
pd.DataFrame(test)
 0                 1
{'a': 1, 'b': 1}  {'a': 2, 'b': 2}
pd.__version__
'1.0.2'

Latest version seems to fix it. However, I can reproduce the faulty behaviour in a fresh conda environment with the old pandas version.

The thread may be closed. However, it should be checked if this problem was intentionally fixed at some point or its fixed by chance.

[simonjayhawkins]

Latest version seems to fix it. However, I can reproduce the faulty behaviour in a fresh conda environment with the old pandas version.

The thread may be closed. However, it should be checked if this problem was intentionally fixed at some point or its fixed by chance.

Thanks @Mereep for the report. I can confirm this issue is a duplicate of #25691 was fixed in #25731

9fbb9e7 is the first new commit
commit 9fbb9e7
Author: Tao He [email protected]
Date: Fri Apr 5 08:45:19 2019 +0800

BUG: Convert tuple to list before `_list_to_arrays` when construct DataFrame. (#25731)

[BenjaminLiuPenrose]

I just try to reproduce the bug.

For pandas<=0.24.2, running pd.DataFrame(tuple([[1],[2]])) will lead to python crash and exit
For pandas>=0.25.0, running pd.DataFrame(tuple([[1],[2]])) is fine as described by @Mereep

[BenjaminLiuPenrose]

In pandas==0.24.2

In pandas==0.25.0

[BenjaminLiuPenrose]

Guess this issue thread can be closed now...