CODEWITHSUNDEEP
pythonregexloops
Johann Schwaller
Johann Schwaller

Reputation: 55

Nested Loop for List in Dictionary

I have a Dictionary that contains lists as values :

{
'List1' : ['Value1', 'Value2', 'Value3'],
'List2' : ['Value1', 'Value2', 'Value3'],
'List3' : ['Value1', 'Value2', 'Value3'],
}

I want to iterate over the values of each list to find regexs, then create a dictionary containing those regexs. That is, for each list of my initial dictionary. Each iteration over my lists (so 3 in the former example) creates 1 row (so 3 rows in total), so I'd then run a code to make a comprehensive unique row.

Not sure if that's clear, but it should look similar to this :

for list in dictionary:
    for value in list:
            column_list_A = []
            if re.search(regex, value):
                column_list_A.append(regex, value).group(1)
            column_list_B = []
            if re.search(regex, value):
                column_list_B.append(regex, value).group(1)
    New_Dictionary = {"column_list_A" : column_list_A, "column_list_B" : column_list_B}
    Df = pd.DataFrame.from_dict(New_Dictionary)
    for column in Df:
        #Code that puts the values of the 3 rows into 1 row

The output should look like this :

      | Column_list_A  |  Column_list_B
----------------------------------------------------
List1 |  match object  | match object  
----------------------------------------------------
List2 |  match object  | match object  
----------------------------------------------------
List3 |  match object  | match object

My questions are :

1) How do I implement the nested for loops ? I've tried using things like iteritems() but it didn't give satisfying results. What exactly should the X and Y be in "for X in Y" for each loop ?

2) Is the indentation correct ?

Upvotes: 1

Views: 557

Answers (2)

Mykola Zotko
Mykola Zotko

Reputation: 17911

If you can use the following dictcomp:

import re
from pprint import pprint

d = {
'List1' : ['Value1', 'Value2', 'Value3'],
'List2' : ['Value1', 'Value2', 'Value3'],
'List3' : ['Value1', 'Value2', 'Value3'],
}

col = ["column_list_A", "column_list_B", "column_list_C"]

def func(a, b, c):
    a = re.match(r'Val(ue\d)', a).group(1)
    b = re.match(r'Valu(e\d)', b).group(1)
    c = re.match(r'Value(\d)', c).group(1)
    return [a, b, c]

new_d = {i: func(*j) for i, *j in zip(col, *d.values())}

pprint(new_d)

Output:

{'column_list_A': ['ue1', 'e1', '1'],
 'column_list_B': ['ue2', 'e2', '2'],
 'column_list_C': ['ue3', 'e3', '3']}

Upvotes: 0

Chilled Geek
Chilled Geek

Reputation: 98

If you want your final output to be a dataframe, I would suggest that you use the panda functions that can handle the looping and regex nicely by themselves without the need for for loops. Here's an example:

import pandas as pd

# read dict in the right orientation
df = pd.DataFrame.from_dict(dictionary, orient="index")

''' # your df will look like this:
>>> df
            0       1       2
List1  Value1  Value2  Value3
List2  Value1  Value2  Value3
List3  Value1  Value2  Value3
'''

# append your regex matches to the dataframe

# e.g. match any of (d,e) followed by a digit
df["match_from_column_0"] = df[0].str.extract(r'([de]\d)')

# e.g. match a digit
df["match_from_column_1"] = df[1].str.extract(r'(\d)')

# save your output as a dataframe
output = df[["match_from_column_0","match_from_column_1"]]

''' # output will look like this:
>>> output
      match_from_column_0 match_from_column_1
List1                  e1                   2
List2                  e1                   2
List3                  e1                   2
'''

# or a dict
output_dict = output.to_dict()
'''
>>> output_dict
{'output1': {'List1': 'e1', 'List2': 'e1', 'List3': 'e1'}, 
'output2': {'List1': 'e2', 'List2': 'e2', 'List3': 'e2'}}
'''

To address your 2 questions:

  • The loop over a dictionary could be something like (assuming for python3): 
    for dict_key, dict_value in dictionary.items():
    # do whatever
  • The loop for over a list could be something like: 
    for value in my_list:
    # do whatever

  • Your lines 3-8 should be outdented (4 spaces from your second for loop indentation)

  • To do it your way (in my opinion the harder way), here's a suggestion (the if statements should need an else clause + append empty string, as they will cause your lists to be of unequal length?):

import re

for key, list_of_values in dictionary.items():
    for value in list_of_values:
        column_list_A = []
        if re.search(regex, value):
            column_list_A.append(re.search(regex, value).group(0))
        else:
            column_list_A.append("")
        column_list_B = []
        if re.search(regex, value):
            column_list_B.append(re.search(regex, value).group(0))
        else:
            column_list_B.append("")
    New_Dictionary = {"column_list_A" : column_list_A, "column_list_B" : column_list_B}
    Df = pd.DataFrame.from_dict(New_Dictionary)
    for column in Df:
        # do your thing

Some references to the documentation:

Hope that helps!

Upvotes: 1

Related Questions