条件过滤

----------------基于numpy

world_alcolhol是numpy的array类型。输入：matrix； 输出：matrix

# Boolean vector corresponding to Canada and 1986.
canada_1986_boolean = (world_alcohol[:,2] == "Canada") & (world_alcohol[:,0] == "1986")
# We can then use canada_1986 to subset a matrix -- it's just a normal boolean vector
print(world_alcohol[canada_1986_boolean,:])

 

---------------基于pandas

• pd.isnull()  输入：dataframe；输出：vector

#dataframe is titanic_survival, the age column has null values(NaN) which need to be excluded while calculation
# age_null is a boolean vector, and has "True" where age is NaN, and "False" where it isn't
age_null = pd.isnull(titanic_survival["age"])
# then use this boolean to filter age column
survival_with_valid_age = titanic_survival["age"][age_null==False]
# do calculation
correct_sum = sum(survival_with_valid_age)
correct_mean_age=correct_sum/len(survival_with_valid_age)

  输入：dataframe；输出：vector

pclass_survival = titanic_survival["fare"][titanic_survival["pclass"]==2]
fare_for_class = pclass_survival.mean()

 输入：dataframe；输出：dataframe

selected_table=table[table['Major_category']==major]

根据列空值过滤行 dropna()

输入：dataframe；输出：dataframe

# do calculation drop rows if certain columns have missing values
new_titanic_survival = titanic_survival.dropna(subset=["age", "body","home.dest"]).reset_index(drop=True)

输入：series；输出：series

#series
series_film=fandango['FILM']
series_rt=fandango['RottenTomatoes']
#use film names as index
series_custom = Series(series_rt.values,index= series_film.values)
#filter
criteria_one = series_custom > 50
criteria_two = series_custom < 75
both_criteria = series_custom[criteria_one & criteria_two]