Pandas rolling function fails with "Wrong number of items passed 2, placement implies 1"-CodePudding

In the following example, I'm trying to calculate the 3-month moving average of values grouped by country using pandas rolling function:

df = pd.DataFrame({
    'country':  ['US','US','US','US', 'US', 'US', 'FR','FR','FR','FR'],
    'year': [1990, 1991, 1992, 1993, 1994, 1995, 1990, 1991, 1992, 1993],
    'gdp': [1.2, 1.4, 1.7, 2.1, 2.3, 1.9, 4.1, 4.6, 4.3, 4.4]
})

print(df)

     country    year    gdp
0    US         1990    1.2
1    US         1991    1.4
2    US         1992    1.7
3    US         1993    2.1
4    US         1994    2.3
5    US         1995    1.9
6    FR         1990    4.1
7    FR         1991    4.6
8    FR         1992    4.3
9    FR         1993    4.4

df['ma'] = df.groupby('country', as_index=False)['gdp'].rolling(3,min_periods=1).mean()

Throws the following error:

ValueError: Wrong number of items passed 2, placement implies 1

Where am I passing 2 items instead of 1?

CodePudding user response：

print(df.groupby('country', as_index=False)['gdp'].rolling(3,min_periods=1).mean())

  country       gdp
6      FR  4.100000
7      FR  4.350000
8      FR  4.333333
9      FR  4.433333
0      US  1.200000
1      US  1.300000
2      US  1.433333
3      US  1.733333
4      US  2.033333
5      US  2.100000

This result returns 2 columns and 1 index. That's why it cannot be assign to 1 column in your df.

Quick fix for problem:

df['ma'] = df.groupby('country', as_index=False)['gdp'].rolling(3,min_periods=1).mean()['gdp']

CodePudding user response：

I don't know about rolling() but I caught an error. It can return 2 Columns in a dataframe but you can provide only one column df['ma']. You can use this to avoid an error.

df[['item1','item2']] = df.groupby('country', as_index=False)['gdp'].rolling(3,min_periods=1).mean()

It can show output like this: