问题描述
程序中断，报错如下AssertionError: 618 columns passed, passed data had 508 columns
Exception has occurred: ValueError
618 columns passed, passed data had 508 columns
AssertionError: 618 columns passed, passed data had 508 columns
The above exception was the direct cause of the following exception:
File "E:\matlab\CHB-MIT-DATA\epilepsy_eeg_classification\preprocessing.py", line 117, in eeg_preprocessing
res = pd.DataFrame(res, columns=column_names)
File "E:\matlab\CHB-MIT-DATA\epilepsy_eeg_classification\preprocessing.py", line 334, in <module>
res = eeg_preprocessing(file, seizures)
ValueError: 618 columns passed, passed data had 508 columns
terminal报错如下：
Backend Qt5Agg is interactive backend. Turning interactive mode on.
弯路
numpy1.19.4改为了numpy1.21.6
pip list： numpy1.21.6
conda list： numpy1.20.1
发现，报错仍旧一样，没有任何改变。
我的怀疑是数据里面有nan值，而程序中并没有处理的方法，是丢弃是补充为0还是补充为平均值。因为如果丢弃的话。
我发现numpy版本不对，卸载的时候，报错说：
PackageNotInstalledError: Package is not installed in prefix于是我使用
conda udate numpy
报错如下：
(base) PS E:\matlab> conda update numpy
Collecting package metadata (repodata.json): done
Solving environment: -
The environment is inconsistent, please check the package plan carefully
The following packages are causing the inconsistency:
- defaults/win-64::anaconda==custom=py37_1
- https://repo.anaconda.com/pkgs/main/win-64::bkcharts==0.2=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::blaze==0.11.3=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::bokeh==0.13.0=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::dask==0.19.1=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::numpydoc==0.8.0=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::odo==0.5.1=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::seaborn==0.9.0=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::sphinx==1.7.9=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::spyder==3.3.1=py37_1
- https://repo.anaconda.com/pkgs/main/win-64::statsmodels==0.9.0=py37h452e1ab_0
- defaults/win-64::_anaconda_depends==5.3.1=py37_0
failed
CondaMemoryError: The conda process ran out of memory. Increase system memory and/or try again.有人说可以直接退到base环境更新conda。
(base) PS E:\matlab> conda update --name base conda
Collecting package metadata (repodata.json): done
Solving environment: |
The environment is inconsistent, please check the package plan carefully
The following packages are causing the inconsistency:
- defaults/win-64::anaconda==custom=py37_1
- https://repo.anaconda.com/pkgs/main/win-64::bkcharts==0.2=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::blaze==0.11.3=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::bokeh==0.13.0=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::dask==0.19.1=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::numpydoc==0.8.0=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::odo==0.5.1=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::seaborn==0.9.0=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::sphinx==1.7.9=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::spyder==3.3.1=py37_1
- https://repo.anaconda.com/pkgs/main/win-64::statsmodels==0.9.0=py37h452e1ab_0
- defaults/win-64::_anaconda_depends==5.3.1=py37_0
failed
CondaMemoryError: The conda process ran out of memory. Increase system memory and/or try again.conda update conda，conda update numpy，conda update --name base conda，在cat（自建虚拟环境）和base里都报错：
CondaMemoryError: The conda process ran out of memory. Increase system memory and/or try again.还有的朋友建议：
conda update conda -c conda-canary但是仍旧不行，报错如下，和前面的报错也一样。
CondaMemoryError：conda 进程内存不足答案 - 爱码网
更新anaconda的版本也报同样的错：
(base) PS E:\matlab> conda update anaconda
Collecting package metadata (repodata.json): done
Solving environment: \
The environment is inconsistent, please check the package plan carefully
The following packages are causing the inconsistency:
- defaults/win-64::anaconda==custom=py37_1
- https://repo.anaconda.com/pkgs/main/win-64::bkcharts==0.2=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::blaze==0.11.3=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::bokeh==0.13.0=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::dask==0.19.1=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::numpydoc==0.8.0=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::odo==0.5.1=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::seaborn==0.9.0=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::sphinx==1.7.9=py37_0
- https://repo.anaconda.com/pkgs/main/win-64::spyder==3.3.1=py37_1
- https://repo.anaconda.com/pkgs/main/win-64::statsmodels==0.9.0=py37h452e1ab_0
- defaults/win-64::_anaconda_depends==5.3.1=py37_0
failed
CondaMemoryError: The conda process ran out of memory. Increase system memory and/or try again.找到好久找到了一个解决方案：
https://github.com/conda/conda/issues/10751
解决方案
说到最后实在是稀松平常，我检查了一下程序逻辑。我发现在传递参数的过程中。
File "E:\matlab\CHB-MIT-DATA\epilepsy_eeg_classification\preprocessing.py", line 117, in eeg_preprocessing
res = pd.DataFrame(res, columns=column_names)
File "E:\matlab\CHB-MIT-DATA\epilepsy_eeg_classification\preprocessing.py", line 334, in <module>
虽然报错在117处，但是，实际上在传递的这两个数据处，通过debug的方式，发现获取的数据的列数要大于实际列数。那是因为，信号的channel变化的，而不是23个固定不变的，所以，只需要把这个固定的23改为，len(channel)就能获取到具体的数字。
将
for i in range(23):
features.extend(eeg_features(temp[i]).tolist())改为
for i in range(len(channels)):
features.extend(eeg_features(temp[i]).tolist())即可。
需要特别说明的是，conda创建的虚拟环境不要胡乱删除，否则会报很多错，你可以根据报错内容删除一些相应的文件，但是不能删除过多的文件。否则会报无数的错误。
这会让你非常头疼。这三篇文章都是我删错文件报错的。
你会发现你无意中删除了一些包的依赖，这就麻烦了。
另外，我的程序报错或许和你的虽然报错一样，但是具体错误的地方不一样，这个时候，你就要好好检查你生成的数据和列名是否符合情况。可以参考这篇文章来看看，或许能够解决你的错误：
已解决ValueError: 4 columns passed, passed data had 2 columns_无 羡ღ的博客-CSDN博客
参考文章