How do I copy a row from an existing dataframe df_a into a new dataframe df_b?

di1a 0
Viewed 40

How do I copy a row from an existing dataframe df_a into a new dataframe df_b? Also, a cell from dataframe df_a into the new dataframe df_b? See the following example:



for index, row in df__data.iterrows():

     for i in range(df__attributes_to_compare.shape[0]):
        
        if row[df__attributes_to_compare["info_a"].values[i]] != row[df__attributes_to_compare["info_b"].values[i]]:      
             
               # new_df.append(key columns values + unsimiliar compared values)

df__data:

key1 key2 key3 attrb1 attrb2 attrb3
sfdg dagd dgsg 12 43 24
afrtf yxcbb ertet 34 45 34

df__attributes_to_compare:

info_a info_b
attrb1 attrb2
attrb1 attrb3

new_df, target output:

key1 key2 key3 value1 value2
sfdg dagd dgsg 12 43
sfdg dagd dgsg 12 24
afrtf yxcbb ertet 34 45

The values that do not match are stored in value1 and value2.

iteration if-statement dataframe pandas python
1 Answers

Here is my solution

import pandas as pd
import numpy as np

data = {"key1":     ["sfdg", "afrtf"],
        "key2":     ["dagd", "yxcbb"],
        "key3":     ["dgsg", "ertet"],
        "attrb1":   ["12", "34"],
        "attrb2":   ["43", "45"],
        "attrb3":   ["24", "34"]}

df__data = pd.DataFrame(data = data)



list_key = ["key1", "key2", "key3"]



# attributes_to_compare = {"info_a": ["attrb1", "attrb1"],
#                          "info_b": ["attrb2", "attrb3"]}

# df__attributes_to_compare = pd.DataFrame(data = attributes_to_compare)



list__attributes_to_compare = ["attrb1", "attrb2", 
                               "attrb1", "attrb3"]



key_set = []
values = []



for index, row in df__data.iterrows():
    
    for i in range(int(len(list__attributes_to_compare)/2)):
        
        i = i*2
        
        if row[list__attributes_to_compare[i]] != row[list__attributes_to_compare[i+1]]:
               
             key_set.append(row.iloc[0:len(list_key)])
             values.append([row[list__attributes_to_compare[i]], row[list__attributes_to_compare[i+1]]])       



df__values = pd.DataFrame(data = values, columns = ["value1", "value2"])

df__key_set = pd.DataFrame(data = key_set, columns = list_key)
df__key_set["new_index"] = np.arange(df__key_set.shape[0])
df__key_set.set_index("new_index", inplace = True)




df__result = df__key_set.merge(df__values, left_index=True, right_index=True)

Does anyone have suggestions for improvement?

di1a0
Related