BUG:TypeError: Cannot interpret 'CategoricalDtype(categories=['no', 'yes'], ordered=False)' as a data type #1692
Description
I just upgraded all my python libraries, and now my previous code is started to fail.
I'm using blaze with pandas. Here is my method code
blaze.data(res)
res contains below data
col1 age ... col31 year
0 yes 55-64 ... NaN 2011
1 no 25-34 ... NaN 2011
2 no 55-64 ... NaN 2011
I'm using below dependencies
- python=3.6.11=h4d41432_2_cpython
- blaze=0.11.3=py36h4e06776_0
- odo=0.5.1=py36h90ed295_0
- pandas=1.0.5=py36h0573a6f_0
- numpy=1.19.1=py36h3849536_2
I'm getting the following error
File "cytoolz/functoolz.pyx", line 667, in cytoolz.functoolz.pipe
File "cytoolz/functoolz.pyx", line 642, in cytoolz.functoolz.c_pipe
File "/home/ubuntu/miniconda/envs/my_env/lib/python3.6/site-packages/blaze/interactive.py", line 153, in data
dshape = discover(data_source)
File "/home/ubuntu/miniconda/envs/my_env/lib/python3.6/site-packages/multipledispatch/dispatcher.py", line 278, in __call__
return func(*args, **kwargs)
File "/home/ubuntu/miniconda/envs/my_env/lib/python3.6/site-packages/odo/backends/pandas.py", line 31, in discover_dataframe
for k in df.columns])
File "/home/ubuntu/miniconda/envs/my_env/lib/python3.6/site-packages/odo/backends/pandas.py", line 31, in <listcomp>
for k in df.columns])
File "/home/ubuntu/miniconda/envs/my_env/lib/python3.6/site-packages/odo/backends/pandas.py", line 23, in dshape_from_pandas
dshape = datashape.CType.from_numpy_dtype(col.dtype)
File "/home/ubuntu/miniconda/envs/my_env/lib/python3.6/site-packages/datashape/coretypes.py", line 779, in from_numpy_dtype
if np.issubdtype(dt, np.datetime64):
File "/home/ubuntu/miniconda/envs/my_env/lib/python3.6/site-packages/numpy/core/numerictypes.py", line 388, in issubdtype
arg1 = dtype(arg1).type
TypeError: Cannot interpret 'CategoricalDtype(categories=['no', 'yes'], ordered=False)' as a data type
It's failing to parse second column values.