[PYTHON] [For recording] Pandas memorandum

This time, I will summarize the events that I did not immediately understand when using Pandas. How do I operate Pandas? I often get lost, so I hope it will be helpful as much as possible.

Create dataset to use

--First, create a data set appropriately. This article will use this simple data as an example.

python

index = ['Product A', 'Product B', 'Product C']
columns = ['spring', 'summer', 'autumn', 'winter']
data = np.array([
    [0,10,20,30],
    [0,0,100,20],
    [50,100,20,40],
])

df = pd.DataFrame(data, index=index, columns=columns)
df.head()

Case (1) Extract and add the column names corresponding to the maximum and minimum values of each row, and their values.

--In other words, for product A, winter is the largest, and for product C, summer is the largest, so insert that season into the new column "Selling season".

python

for index, row in df.iterrows():
    
#     df.ix[index, 'Selling season'] = row.argmax()
    df.ix[index, 'Selling season'] = row.idxmax() #You can do either

#     df.ix[index, 'Unsellable season'] = row.argmin()
    df.ix[index, 'Unsellable season'] = row.idxmin() #You can do either
    
    df.ix[index, 'MAX'] = row.max()
    df.ix[index, 'MIN'] = row.min()

df.head()

Case (2) Extract and add the column name that becomes a positive integer first in each row

――In other words, I want to put out the season when each product starts to sell for the first time. A is summer, B is autumn

python

for index, row in df.iterrows():
#index[0]Take out the first one that meets the conditions with
    df.ix[index, 'Season to start selling'] = row[row > 0].index[0] 
    
df.head()

Case ③ I want to extract a line that does not contain any 0

――In other words, I want to extract the line of product C that sells all year round.

python

#not_zero_df = df.query('spring>0 and summer>0 and autumn>0 and winter> 0')
#not_zero_df = df.query('spring!=0 and summer!=0 and autumn!=0 and winter!= 0')

#↑ It's okay, but it can be outside as a variable. Below "@Just add
hoge1,hoge2,hoge3,hoge4 = 0,0,0,0 
not_zero_df = df.query('spring> @hoge1 and summer> @hoge2 and autumn> @hoge3 and winter> @hoge4')
    
not_zero_df.head()

--By the way, if you want to extract something other than 0 by focusing only on summer, you can also write like this (extract lines where summer is not 0).

python

hoge = 0 
not_zero_df = df.query('summer!= @hoge')
    
not_zero_df.head()

** We will continue to add it as needed **