BUG: ExtensionArrays whose elements are non-numeric numpy arrays crash Series.repr() (original) (raw)

Code Sample, a copy-pastable example

Prerequisites:

$ pip install pandas numpy memoized-property $ pip install git+https://github.com/frreiss/text-extensions-for-pandas

Reproduce the problem, using TensorArray, our extension type for storing tensors in a Pandas series:

import pandas as pd import numpy as np import text_extensions_for_pandas as tp

Integers work

int_tensors = np.array([[1, 2], [3, 4]])
int_tensor_series = pd.Series(tp.TensorArray(int_tensors))
int_tensor_series

0 [1 2] 1 [3 4] dtype: TensorType

Boolean values don't work

bool_tensors = np.array([[True, False], [False, True]])
bool_tensor_series = pd.Series(tp.TensorArray(bool_tensors))
bool_tensor_series

ValueError Traceback (most recent call last) ~/pd/covid-notebooks/env/lib/python3.7/site-packages/IPython/core/formatters.py in call(self, obj) 700 type_pprinters=self.type_printers, 701 deferred_pprinters=self.deferred_printers) --> 702 printer.pretty(obj) 703 printer.flush() 704 return stream.getvalue()

[...many lines of stack trace...]

~/pd/covid-notebooks/env/lib/python3.7/site-packages/pandas/io/formats/format.py in _format_strings(self) 1255 fmt_values = [] 1256 for i, v in enumerate(vals): -> 1257 if not is_float_type[i] and leading_space: 1258 fmt_values.append(" {v}".format(v=_format(v))) 1259 elif is_float_type[i]:

ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

Problem description

Series.__repr__() invokes ExtensionArrayFormatter to render extension types. If the individual elements managed by an ExtensionArray are numpy arrays (or slices of a larger numpy array), then ExtensionArrayFormatter uses Pandas' facilities for rendering numpy arrays. These facilities comprise the base class GenericArrayFormatter and subclasses such as FloatArrayFormatter for handling specific types (see pandas/io/formats/format.py). The formatters for numeric types can render numpy arrays with more than one dimension, but the base class GenericArrayFormatter cannot. The limitation appears to be an oversight. I will submit a small pull request with a fix in a few minutes.

Expected Output

The above code should output something like this:

0 [True False] 1 [False True] dtype: TensorType

Output of pd.show_versions()

Details

INSTALLED VERSIONS

commit : None
python : 3.7.7.final.0
python-bits : 64
OS : Darwin
OS-release : 19.4.0
machine : x86_64
processor : i386
byteorder : little
LC_ALL : None
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8

pandas : 1.0.3
numpy : 1.18.1
pytz : 2019.3
dateutil : 2.8.1
pip : 20.0.2
setuptools : 46.1.3.post20200330
Cython : 0.29.15
pytest : None
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : 2.11.1
IPython : 7.13.0
pandas_datareader: None
bs4 : None
bottleneck : None
fastparquet : 0.3.3
gcsfs : None
lxml.etree : None
matplotlib : 3.1.3
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : 0.13.0
pytables : None
pytest : None
pyxlsb : None
s3fs : None
scipy : 1.4.1
sqlalchemy : None
tables : None
tabulate : 0.8.3
xarray : None
xlrd : None
xlwt : None
xlsxwriter : None
numba : 0.49.0