Removing a column that is not in the columnset is error

[ztsweet]

import datatable as dt
from datatable import f
#
#print(dt.__version__)
# '1.0.0'
df1 = dt.Frame({'a':[1, 2, 3], 'b':[2, 3, 4],'c':[1., 2., 3.]})
df1[:, f[:].remove(f['d'])]
#
# KeyError: Column d does not exist in the Frame; did you mean a, b or c?

[samukweku]

@ztsweet , the error message seems right. What is your expectation?

[ztsweet]

document/user guide/f-expressions

[samukweku]

@ztsweet neat… @st-pasha @oleksiyskononenko can weigh in on this … seems like a bug

[samukweku]

@ztsweet kindly add the datatable version. Edit it using the guidelines, so the maintainers can properly look into it

[ztsweet]

@samukweku the datatable's version is '1.0.0'

[st-pasha]

So, either the behavior is wrong here (i.e. there should not be an exception), or the documentation is wrong.

But which one? What are your guys' thoughts on this?

[pradkrish]

Throwing an error like it's done now seems more appropriate to me. I suggest updating the documentation to match the behaviour.

[samukweku]

The documentation seems the right way to go, if we are treating it like a set

[pradkrish]

@oleksiyskononenko what are your thoughts on this?

[oleksiyskononenko]

@pradkrish The Python's .remove() method results in an error for both lists and sets, when the element to be removed is absent in the list/set.

l = ["a", "b", "c"]
l.remove("d") #ValueError: list.remove(x): x not in list

s = {"a", "b", "c"}
s.remove("d") #KeyError: 'd'

So my feeling is that we need to adjust documentation.