CODEWITHSUNDEEP
pythonpandasif-statement
Kben59
Kben59

Reputation: 388

IF statement : how to print names of columns in Python

Here is my issue

I have this DataFrame :

df = pd.DataFrame({'Date':[1,2,3],
                   'Paul':[8,10,13],
                   'Mathieu':[18,20,2],
                   'Jacques':[2,1,70]})

df = df.set_index('Date')

My goal is to create an IF statement with those conditions : IF the last value of the rolling mean 2 days is < the rolling mean 3 days and the fist value (Day 1) is > the last value (Day 3) then print the name of the column.

This is what I started : 

def test(data):
    end = data.iloc[-1]
    start = data.iloc[0]
    end_rolling_2D = data.rolling(2).mean().iloc[-1]
    end_rolling_3D = data.rolling(3).mean().iloc[-1]

    if end_rolling_2D < end_rolling_3D and start > end :
        print(data.columns)

But I have this Error :

ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().

And I don't really know how to solve that.I know that the aswer should be 'Mathieu' only because he's the only one that met the conditions.

I am very new to Python, so if anyone has an idea to solve that, you are welcome!

Thank you.

Upvotes: 0

Views: 256

Answers (2)

Ch3steR
Ch3steR

Reputation: 20669

start, end, end_rolling_2D and end_rolling_3D are Series object, use pd.Series.all()

Replace

if end_rolling_2D > end_rolling_3D and start > end

with

if ((end_rolling_2D>end_rolling_3D) & (start>end)).all()

To get the location where the condition is satisfied use boolean indexing

mask = (end_rolling_2D<end_rolling_3D) & (start>end)
print(df.columns[mask])
# Index(['Mathieu'], dtype='object')

Upvotes: 2

Vinod Kumar
Vinod Kumar

Reputation: 1622

Change your function to:

def test(data):
    end = data.iloc[-1]
    start = data.iloc[0]
    end_rolling_2D = data.rolling(2).mean().iloc[-1]
    end_rolling_3D = data.rolling(3).mean().iloc[-1]

    if ((end_rolling_2D > end_rolling_3D) & (start > end)).all() :
        print(data.columns)

Please note that I have added brackets. And replaced and by & However, I do not expect any data which satisfies your criteria.

Upvotes: 0

Related Questions