Question

这里总共有熊猫新手，所以请多多怜悯。我有一个数据样本，上面粘贴了形状的年度条目：

{"Country":{"0":"Italy","1":"Italy","2":"Italy","3":"Italy","4":"Italy","5":"Italy","6":"Italy","7":"France","8":"France","9":"France","10":"France","11":"France","12":"France","13":"Spain","14":"Spain","15":"Spain","16":"Spain","17":"Spain","18":"Spain","19":"Spain"},"Year":{"0":2004,"1":2005,"2":2006,"3":2007,"4":2008,"5":2009,"6":2010,"7":2006,"8":2007,"9":2008,"10":2009,"11":2010,"12":2011,"13":2007,"14":2008,"15":2009,"16":2010,"17":2011,"18":2012,"19":2013},"Revenue":{"0":1000,"1":1200,"2":1300,"3":1400,"4":1450,"5":1300,"6":1200,"7":2200,"8":2100,"9":1900,"10":2300,"11":2400,"12":2500,"13":1150,"14":1230,"15":1300,"16":1200,"17":1050,"18":900,"19":950}}

我需要一种方法来仅过滤所有国家/地区的常见年份，例如2007、2008、2009和2010。

我认为我应该制定一个公式并将其应用，但是我似乎找不到自己的出路。

Answer 1

使用两次nunique：获得唯一国家的数量n，然后过滤年份，仅将唯一国家的数量等于n

n = df.Country.nunique()
s = df.groupby('Year').Country.nunique().eq(n)

>>> print(s)
Year
2004    False
2005    False
2006    False
2007     True
2008     True
2009     True
2010     True
2011    False
2012    False
2013    False
Name: Country, dtype: bool

要获得年份，

>>> print(s[s].index)
[2007, 2008, 2009, 2010]

也可以使用set intersection

>>> set.intersection(*df.groupby('Country').Year.agg(set))
{2007, 2008, 2009, 2010}

Answer 2

选项1
pivot + dropna

df.pivot('Year', 'Country', 'Revenue').dropna().index

选项2
crosstab + all

u = pd.crosstab(df.Year, df.Country)
u[u.all(1)].index

两种产品：

Int64Index([2007, 2008, 2009, 2010], dtype='int64', name='Year')

根据合并条件过滤熊猫中的行

2 个答案: