假设您使用Pandas数据框来保存数据，则可以获得一组唯一值，然后对它们进行迭代以确定它们在每列中的占用情况。

import pandas as pd

df = pd.read_csv('path/to/file.csv')

def in_col_fraction(df, value):
    in_cols = [(value in df[c]) for c in df]
    return sum(in_cols) / len(in_cols)

# this holds the column presence of all items in your data
presence_dict = {
    value: in_col_fraction(df, value)
    for value in df.stack().unique()
}

# this is the list of just the items that have 95% occupancy or higher
over_95_list = [v for v in df.stack().unique() if (in_col_fraction(df, v) >= 0.95)]

这是一个R解决方案，它使用了一个有四列的示例 Dataframe ，并且找到了至少75%的列中出现的值。

all_vals <- unique(unlist(dat))

pct_w_val <- sapply(
  all_vals, 
  \(val) mean(sapply(dat, \(col) val %in% col))
)

all_vals[pct_w_val >= .75]
# [1] "x" "e" "p" "v" "s" "q" "o" "w" "n"

• 示例数据：*

x一个一个一个一个x一个一个二个x