pandas 检查 MultiIndex 中的一个特定索引中是否存在值 dataframe

Question

I have a MultiIndex data frame called df with 3 indexes (Fruit, Color, Taste). I want to search 1 specific index , that index being Color and see if the value exists in it.

For example: the code would look something like this. Color is an index in the dataframe not just a column.

if 'purple' in 'Color':
    print('yes')
else:
    print('no')

I only want it to search the Color not any other indexes/columns

                       Quantity     Quality
Fruit   Color   Taste
apple   red     tart     12          good
lemon   yellow  sour     11          average
grapes  purple  sweet     5          bad
lime    green   citrus    3          excellent

Thank you so much for your time!

Answer 1

Try xs to grab a cross-section from the DataFrame:

df.xs('purple', level=1, drop_level=False)

                     Quantity Quality
Fruit  Color  Type                   
grapes purple sweet         5     bad

Just to test if it exists or not check if the cross-section is empty or not:

df.xs('purple', level=1).empty

if not df.xs('purple', level=1).empty:
    print('yes')
else:
    print('no')

---

Alternatively reset_index and use any to see if there are any matches:

df.reset_index('Color')['Color'].eq('purple').any()

True

if df.reset_index('Color')['Color'].eq('purple').any():
    print('yes')
else:
    print('no')

Answer 2

You can use this if you want a tabular output:

def check(data:pd.DataFrame,l:list):
    c = data.index.get_level_values("Color").isin(l)
    return np.where(c,'yes','no')

df['Result'] = check(df,['purple'])

---

print(df)

                       Quantity    Quality Result
Fruit  Color  Type                              
apple  red    tart          12       good     no
lemon  yellow sour          11    average     no
grapes purple sweet          5        bad    yes
lime   green  citrus         3  excellent     no

Answer 3

You can use get_level_values to filter.

if "purple" in df.index.get_level_values('Color'):
    print('yes')
else:
    print('no')

Answer 4

You can also check if the value is in index.unique(level) . This way the whole index would not need to be copied into a list. It might also make the subsequent search faster.

if "purple" in df.index.unique(level="Color"): 
    print("yes")
else: 
    print("no")