I encountered the following bugs in the python version of MonetDB Lite (I installed with pip install):
In Python 3 creating a table from a Pandas dataframe doesn't work at all, because all columns in the data frame are of type object so an exception is raised at create.
In Python 2 Unicode characters in a CSV raise the exception:
DatabaseError: pyapi.eval:Invalid string encoding used. Please return a regular ASCII string, or a Numpy_Unicode object.
I am attaching a small testcase for both, just comment/uncomment the correct line in the load.py
PS: another thing, although it's not exactly a bug is that pip install monetdblite fails if numpy is not installed. Maybe consider adding it as a dependency? (although I do remember the discussion we had with @hannesmuehleisen a few days ago :) )
Hello guys,
I encountered the following bugs in the python version of MonetDB Lite (I installed with
pip install
):In Python 3 creating a table from a Pandas dataframe doesn't work at all, because all columns in the data frame are of type
object
so an exception is raised atcreate
.In Python 2 Unicode characters in a CSV raise the exception:
DatabaseError: pyapi.eval:Invalid string encoding used. Please return a regular ASCII string, or a Numpy_Unicode object.
I am attaching a small testcase for both, just comment/uncomment the correct line in the load.py
Library versions:
python2 2.7.13 python3 3.5.3 monetdblite 0.1.9 numpy 1.13.1 pandas 0.20.3
cc @Mytherin
mdblite_python_bugreport.tar.gz
Panos.
PS: another thing, although it's not exactly a bug is that
pip install monetdblite
fails if numpy is not installed. Maybe consider adding it as a dependency? (although I do remember the discussion we had with @hannesmuehleisen a few days ago :) )