There seem to be a bug in pyeviews. When a dataframe is long enough, it's not possible to push it into Eviews using "PutPythonAsWF".
Below is some reproducible code that has also been posted on Github.
Code: Select all
import pandas as pd
import numpy as np
ind = pd.date_range('2016-01-01 00:00:00', periods = 2000, freq = 'T')
x = pd.Series(np.random.randn(2000), index = ind, name = 'x')
y = pd.Series(np.random.randn(2000), index = ind, name = 'y')
z = pd.Series(np.random.randn(2000), index = ind, name = 'z')
df = pd.concat([x,y,z], axis = 1)
import pyeviews as evp
evapp = evp.GetEViewsApp(instance = 'new', showwindow = True)
evp.PutPythonAsWF(df, app = evapp)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "C:\Users\bjorn.backgard\AppData\Local\Continuum\Anaconda2\lib\site-packa
ges\pyeviews\__init__.py", line 165, in PutPythonAsWF
app.Run(create)
_ctypes.COMError: (-2147221424, None, (u'Incomplete command in batch mode in "CR
EATE MIN(5-6, 0:0:0-23:59:0) 1/1/2016 00:00:00 1/2/2016 09:19:00".', u'EViews.Ap
plication.9.Run', None, 0, None))
df = df[1:500]
evp.PutPythonAsWF(df, app = evapp)
The last line will execute without error, so the issue definitely has to do with the length of the dataset. It doesn't change matters if two of the three columns are deleted.
Is it possible to get the developer involved, or to at least get to know whether the module is still being developed? I noticed no changes has been made during the last six months.
Best regards,
Björn Backgård