pandas filter by two columns (python)

问题

I have a pandas DataFrame (df) with many columns, two of which are "Year" and "col_1"

I also have a extraction criteria summarised in a list(Criteria):

[1234,5432,...,54353,654,1234].

I would like to extract the subset of this DataFrame if the following criteria are met:

((df.Year==1990) & (df.col_1>=Criteria[0])) or

((df.Year==1991) & (df.col_1>=Criteria[1])) or

((df.Year==1992) & (df.col_1>=Criteria[2])) or 

...

((df.Year==2010) & (df.col_1>=Criteria[20])) or

((df.Year==2011) & (df.col_1>=Criteria[21]))

Although I can list out all the combination of these criteria, I would like to do this in one short line, something like:

df = df[df[['col_1','col_2']].apply(lambda x: f(*x), axis=1)]

(from how do you filter pandas dataframes by multiple columns)

Please advise how I can do it. Thank you.

回答1:

Sample DataFrame:

df = pd.DataFrame({'col_1':[2000,1,54353,5],
                   'Year':[1990,1991,1992,1993],
                   'a':range(4)})

print (df)
   col_1  Year  a
0   2000  1990  0
1      1  1991  1
2  54353  1992  2
3      5  1993  3

Create helper dictionary by criteria and years combinations:

Criteria = [1234,5432,54353,654,1234]
years = np.arange(1990, 1990 + len(Criteria))
d = dict(zip(years, Criteria))
print (d)
{1990: 1234, 1991: 5432, 1992: 54353, 1993: 654, 1994: 1234}

Last map by column year and filter by boolean indexing:

df = df[df['col_1'] >= df['Year'].map(d)]
print (df)
   col_1  Year  a
0   2000  1990  0
2  54353  1992  2

Detail:

print (df['Year'].map(d))
0     1234
1     5432
2    54353
3      654
Name: Year, dtype: int64

print (df['col_1'] >= df['Year'].map(d))

0     True
1    False
2     True
3    False
dtype: bool

来源：https://stackoverflow.com/questions/51414814/pandas-filter-by-two-columns-python

标签

python

pandas

filter

multiple-columns