Delete 90% of random rows by condition pandas

Question

I have a pandas dataframe and want to delete 90% of data which satisfies condition.

The condition is very simple. If the value of the column "Parameter1" is greater than a threshold, then delete it.

My question is how to delete 90% of them, not 90% values in a row, but random

Answer 1

Use boolean indexing with sample:

df = pd.DataFrame({
    'A': [5] * 20 + [1] * 2,
    'B': list(range(22))
})

df = df.drop(df[df['A'] > 4].sample(frac=.9).index)
print (df)
    A   B
11  5  11
15  5  15
20  1  20
21  1  21