FutureWarning: Series.base is deprecated and will be removed in a future version

[justincart]

I am using python3.7.2 with xgboost 0.81 and got a lot of the following error messages:

/Users/jianhuitian/Documents/Programs/Anaconda3/anaconda3/lib/python3.7/site-packages/xgboost/core.py:587: FutureWarning: Series.base is deprecated and will be removed in a future version
if getattr(data, ‘base’, None) is not None

How to get rid of this annoying issue?

[thvasilo]

I did a quick grep of the repo and cannot find the offending line if getattr(data, ‘base’, None) is not None, or any relevant reference to Series.base which originates from Pandas.

Not sure where why this warning pops up.

[jrathgeber]

Same issue

[astrowonk]

Getting the same warning.

The code is here:

        if getattr(data, 'base', None) is not None and \
           data.base is not None and isinstance(data, np.ndarray) \
           and isinstance(data.base, np.ndarray) and (not data.flags.c_contiguous):
            self.set_float_info_npy2d(field, data)
            return

From what I can tell about the history of the .base method it's a relic of deeper connections to numpy. I think simply replacing data.base with data.values.base would eliminated the warning and get the same result.

[hdinh]

Here's the code that produces this for me, it seems to be a combination of passing a pandas Series as the sample_weight.

import pandas as pd
import numpy as np
from xgboost import XGBClassifier

X = np.array([[0], [0], [1]])
y = np.array([0, 0, 1])
w = pd.Series([1, 1, 1])
clf = XGBClassifier()
clf.fit(X, y, w)

Looks like it's happening for pandas>=0.23.0 because the pandas Series.base property has been deprecated:
https://github.com/pandas-dev/pandas/blob/master/pandas/core/base.py#L779
from pandas-dev/pandas#20419

[yangjx12]

A simple way, ignore the warning.

import warnings
warnings.simplefilter(action='ignore', category=FutureWarning)

[H4dr1en]

A simple way, ignore the warning.

import warnings
warnings.simplefilter(action='ignore', category=FutureWarning)

Ignoring warnings should never be a solution

[trivialfis]

Suggestions are welcomed. :-)

[hcho3]

#4337 should eliminate FutureWarning

[songyang-dev]

Hi there, I am facing this same issue right now with the latest stable release of xgboost (0.9). I'm using an older version of python than the one used by OP. I have 3.6.8.

There is another warning message that keeps popping up. I will create a different issue for this.

EDIT: The other warning can be suppressed by doing what it is asking. That's what was suggested on another issue.

[hcho3]

@XsongyangX 0.90 version was released on May 20, and the fix was made on July 4. So you'll see see the warning in 0.90. The next release (1.0.0) will not have the warning.

[trivialfis]

Sorry, wrong click.