pandas 熊猫平均函数的 NaN 结果
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/18173873/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
NaN results with pandas mean function
提问by abuteau
I try to have the mean of a row in my Python DataFrame, but I get a NaN return for every row. Why I get this result and how can I solve that ?
我尝试在我的 Python DataFrame 中获得一行的平均值，但我得到每一行的 NaN 返回值。为什么我得到这个结果，我该如何解决？
Goog key ratios : http://www.gogofile.com/Default.aspx?p=sc&ID=635118193040317500_6234
Goog 关键比率：http: //www.gogofile.com/Default.aspx?p=sc&ID=635118193040317500_6234
path = 'GOOG Key Ratios.csv'
#print(open(path).read())
data = pd.read_csv(path, skiprows = 2, names = ['Y0','Y1','Y2','Y3','Y4','Y5','Y6','Y7','Y8','Y9','Y10'], index_col = 0)
noTTM = data.iloc[:,0:10]
print(data.mean(1))
grossMargin = noTTM[2:3]
print(grossMargin.mean(1))
Return:
返回：
Gross Margin % NaN
dtype: float64
Regards,
问候，
回答by Phillip Cloud
The reason you have a bunch of nanvalues is because you don't have homogeneous column types. So, for example when you try to average across the columns it doesn't make sense because pandas.read_csvwill only convert into a numeric column if it makes sense, e.g., you don't have string dates or other text in the same column as numbers.
您有一堆nan值的原因是因为您没有同类的列类型。因此，例如，当您尝试跨列求平均值时，这是没有意义的，因为pandas.read_csv只有在有意义的情况下才会转换为数字列，例如，您在同一列中没有字符串日期或其他文本作为数字。
I recommend also that you do a simple df.head()to check your data before doing even simple analyses. It will save you a lot of time in the future when you're wondering why your output is "weird".
我还建议df.head()您在进行简单的分析之前先检查一下您的数据。当您想知道为什么您的输出“奇怪”时，它将为您节省大量时间。
That said, you can do the following to convert things to numeric values, but this isn't necessarily guaranteed to make sense:
也就是说，您可以执行以下操作将事物转换为数值，但这不一定保证有意义：
In [35]: df = read_csv('GOOG Key Ratios.csv', skiprows=2, index_col=0, names=['Y%d' % i for i in range(11)])
In [36]: df.head() # not homogeneously typed columns
Out[36]:
Y0 Y1 Y2 Y3 Y4 \
NaN 2003-12 2004-12 2005-12 2006-12 2007-12
Revenue USD Mil 1,466 3,189 6,139 10,605 16,594
Gross Margin % 57.3 54.3 58.1 60.2 59.9
Operating Income USD Mil 342 640 2,017 3,550 5,084
Operating Margin % 23.4 20.1 32.9 33.5 30.6
Y5 Y6 Y7 Y8 Y9 Y10
NaN 2008-12 2009-12 2010-12 2011-12 2012-12 TTM
Revenue USD Mil 21,796 23,651 29,321 37,905 50,175 55,797
Gross Margin % 60.4 62.6 64.5 65.2 58.9 56.7
Operating Income USD Mil 6,632 8,312 10,381 11,742 12,760 12,734
Operating Margin % 30.4 35.1 35.4 31.0 25.4 22.8
In [37]: df.convert_objects(convert_numeric=True).head()
Out[37]:
Y0 Y1 Y2 Y3 Y4 Y5 Y6 Y7 Y8 Y9 Y10
NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
Revenue USD Mil NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
Gross Margin % 57.3 54.3 58.1 60.2 59.9 60.4 62.6 64.5 65.2 58.9 56.7
Operating Income USD Mil 342.0 640.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN
Operating Margin % 23.4 20.1 32.9 33.5 30.6 30.4 35.1 35.4 31.0 25.4 22.8
