pandas 掩码 df 分配中的 KeyError

Question

我继承了一个代码库，它非常依赖 DataFrame.assign 和 dict 解包参数，这不是我以前见过的很多东西。

我今天做了一些测试，我想我一定遇到了边缘情况，我一直在寻找解释and/or过去几个小时的解决方案。

我无法共享数据，但我已经成功创建了以下 MRE。

import sys
import pandas as pd

print("python", sys.version)
print("pandas", pd.__version__)

# template df for the output format
def template_df():
    return pd.DataFrame(
        columns=["id", "subid", "empty", "dat1", "dat2", "dat3", "dat4"]
    )

# input data looks like this
df = pd.DataFrame({
    "id1": [None, None, "0039"],
    "id2": ["10", "12", "a1"],
    "dat": [601, 482, 890],
})

# filter on id2 like 'a%'
m1 = df["id2"].str.startswith("a")

# start building output with input data and constant data
output = template_df().assign(**{
    "id": df.loc[m1, "id2"],
    "subid": df.loc[m1, "id1"],
    "dat1": df.loc[m1, "dat"],
    "dat2": "constant2",
})

# filter for id1 = '0039'
m2 = output["subid"].str.match("0039")

# add data for id1 = '0039' only
output[m2] = output[m2].assign(**{"dat3": "dependent3", "dat4": "dependent4"})

当我执行上面的代码时，我得到：

% python soq.py
python 3.9.12 (main, Mar 26 2022, 15:51:15)
[Clang 13.1.6 (clang-1316.0.21.2)]
pandas 1.4.1
Traceback (most recent call last):
  File "/path/to/src/dir/soq.py", line 35, in <module>
    output[m2] = output[m2].assign(**{"dat3": "dependent3", "dat4": "dependent4"})
  File "/path/to/src/dir/.venv/lib/python3.9/site-packages/pandas/core/frame.py", line 3643, in __setitem__
    self._setitem_array(key, value)
  File "/path/to/src/dir/.venv/lib/python3.9/site-packages/pandas/core/frame.py", line 3678, in _setitem_array
    self.iloc[indexer] = value
  File "/path/to/src/dir/.venv/lib/python3.9/site-packages/pandas/core/indexing.py", line 716, in __setitem__
    iloc._setitem_with_indexer(indexer, value, self.name)
  File "/path/to/src/dir/.venv/lib/python3.9/site-packages/pandas/core/indexing.py", line 1688, in _setitem_with_indexer
    self._setitem_with_indexer_split_path(indexer, value, name)
  File "/path/to/src/dir/.venv/lib/python3.9/site-packages/pandas/core/indexing.py", line 1724, in _setitem_with_indexer_split_path
    self._setitem_with_indexer_frame_value(indexer, value, name)
  File "/path/to/src/dir/.venv/lib/python3.9/site-packages/pandas/core/indexing.py", line 1813, in _setitem_with_indexer_frame_value
    self._setitem_single_column(loc, val, pi)
  File "/path/to/src/dir/.venv/lib/python3.9/site-packages/pandas/core/indexing.py", line 1875, in _setitem_single_column
    ser = value[np.argsort(pi)]
  File "/path/to/src/dir/.venv/lib/python3.9/site-packages/pandas/core/series.py", line 984, in __getitem__
    return self._get_with(key)
  File "/path/to/src/dir/.venv/lib/python3.9/site-packages/pandas/core/series.py", line 1019, in _get_with
    return self.loc[key]
  File "/path/to/src/dir/.venv/lib/python3.9/site-packages/pandas/core/indexing.py", line 967, in __getitem__
    return self._getitem_axis(maybe_callable, axis=axis)
  File "/path/to/src/dir/.venv/lib/python3.9/site-packages/pandas/core/indexing.py", line 1191, in _getitem_axis
    return self._getitem_iterable(key, axis=axis)
  File "/path/to/src/dir/.venv/lib/python3.9/site-packages/pandas/core/indexing.py", line 1132, in _getitem_iterable
    keyarr, indexer = self._get_listlike_indexer(key, axis)
  File "/path/to/src/dir/.venv/lib/python3.9/site-packages/pandas/core/indexing.py", line 1327, in _get_listlike_indexer
    keyarr, indexer = ax._get_indexer_strict(key, axis_name)
  File "/path/to/src/dir/.venv/lib/python3.9/site-packages/pandas/core/indexes/base.py", line 5782, in _get_indexer_strict
    self._raise_if_missing(keyarr, indexer, axis_name)
  File "/path/to/src/dir/.venv/lib/python3.9/site-packages/pandas/core/indexes/base.py", line 5842, in _raise_if_missing
    raise KeyError(f"None of [{key}] are in the [{axis_name}]")
KeyError: "None of [Int64Index([0], dtype='int64')] are in the [index]"

Answer 1

您没有正确使用 assign()。您应该改用 .loc:

output.loc[m2, ['dat3', 'dat4']] = ["dependent3", "dependent4"]

有关详细信息，请参阅 this。

pandas 掩码 df 分配中的 KeyError

KeyError in pandas assign of masked df

python

dataframe

python-3.x

pandas