My dataset looks as follows:

I have a total of 60 years of data under consideration, with each country able to contain a max of one row per crop per year- to a total of 240 rows per year. I have used the .value_counts function to evaluate which countries are not whole and would like to create a subset containing only the countries that have a complete set of data (ie. where the .value_counts function returned a value of 240 for that country).

I have tried doing this as follows: df_240 = df.loc[df["Area"].value_counts == 240] however I am having no luck.

The error I am recieving is:

KeyError: 'False: boolean label can not be used without a boolean index'

CodePudding user response：

This is what you are looking for:

df1 = df.groupby(["Area", "Year"]).filter(lambda x: len(x) == 240)

CodePudding user response：

Here is grouping by 'Area' and fetching 'Year' data is passed to transform where 'count' is used for counting. It turns out a logical mask, which is substituted for selecting rows. Indexes are dropped and drop=True is applied.

import pandas as pd

ind = df.groupby('Area')['Year'].transform('count') >= 240
df1 = df[ind].reset_index(drop=True)

print(df1)