[Tensorflow 2.0] TypeError: '(slice(None, None, None), slice(None, -1, None))' is an invalid key

/usr/local/lib/python3.7/dist-packages/pandas/core/indexes/base.py in get_loc(self, key, method, tolerance)
   2896             casted_key = self._maybe_cast_indexer(key)
   2897             try:
-> 2898                 return self._engine.get_loc(casted_key)
   2899             except KeyError as err:
   2900                 raise KeyError(key) from err

pandas/_libs/index.pyx in pandas._libs.index.IndexEngine.get_loc()

pandas/_libs/index.pyx in pandas._libs.index.IndexEngine.get_loc()

TypeError: '(slice(0, 7, None), slice(None, -1, None))' is an invalid key

 

RNN 계열의 Time Series 딥러닝을 하다보면 발생하는 문제입니다.

저도 이거해결하느라 삽질을 좀 많이했는데요.. 원인은 이렇습니다.

 

Pandas df(dataframe)으로 불러온 데이터는 element들에 접근할시 x[0:1] 형태보다 x.iloc[0:1]의 형태를 고수해주어야합니다. 

물론, x[0:1] 이런식으로도 접근은 할 수 있지만, 상기와 같은 pandas 접근에러가 뜰때는 정석으로 iloc을 붙여주는걸 추천드립니다.

 

#Data Windowing

def sliding_window(time_series, seq_length):

    dataX = []

    dataY = []

    for i in range(0, len(time_series) - seq_length): 

        print(i)

        _x = time_series.iloc[i:i + seq_length, : -1 ]

        _y = time_series.iloc[i + seq_length, [-1]]  

        print(_x, "->", _y)

        dataX.append(_x)

        dataY.append(_y)

    return np.array(dataX), np.array(dataY)

trainX, trainY = sliding_window(trainSet, seqLength)

testX, testY = sliding_window(testSet, seqLength)

https://stackoverflow.com/questions/55291667/getting-typeerror-slicenone-none-none-0-is-an-invalid-key

Getting TypeError: '(slice(None, None, None), 0)' is an invalid key

 

[Errorcode]

pandas/_libs/index.pyx in pandas._libs.index.IndexEngine.get_loc()

TypeError: '(slice(None, None, None), slice(None, -1, None))' is an invalid key

/usr/local/lib/python3.7/dist-packages/pandas/core/indexes/base.py in get_loc(self, key, method, tolerance) 2896             casted_key = self._maybe_cast_indexer(key) 2897             try: -> 2898                 return self._engine.get_loc(casted_key) 2899             except KeyError as err: 2900                 raise KeyError(key) from err