Python字典值到键和附加其他键

Question

I have a python DataFrame with the following:

myDF = pd.DataFrame({"COLUMN_NAME": ["Col1", "Col2", "Col3", "Col4"],
"RULE_1": ["NULL", "DUPLICATE", "TEXT-ONLY", "INTEGER-ONLY"],
"RULE_2": ["DUPLICATE", np.nan, "DUPLICATE", np.nan] })

How can I convert this into a dictionary that looks like that:

my_dict = {"Col1": ["NULL", "DUPLICATE"], "Col2": ["DUPLICATE"], "Col3": ["TEXT-ONLY", "DUPLICATE"], "Col4": ["INTEGER-ONLY"]}

I am stuck doing multiple loops but not really finding a solution.

final_rules_dict = defaultdict(list)
for k in rule_dict:
    row_dict = rule_dict[k]
    for k in row_dict:
        col_name = k
        if col_name == "COLUMN_NAMES":
            final_rules_dict[col_name].append()

Answer 1

Here is a solution that works with the format you have given us

import json
myDF = pd.DataFrame({"COLUMN_NAME": ["Col1", "Col2", "Col3", "Col4"],
"RULE_1": ["NULL", "DUPLICATE", "TEXT-ONLY", "INTEGER-ONLY"],
"RULE_2": ["DUPLICATE", np.nan, "DUPLICATE", np.nan] })

myDF = myDF.T.reset_index(drop=True)
myDF.columns = myDF.iloc[0]
{
    col: [x for x in rows if x is not None] for col, rows in json.loads(pd.io.json.dumps(myDF[1:].to_dict(orient='list'))).items()
}

Answer 2

You can prepare your empty dict:

my_dict = {"Col1": [], "Col2": [], "Col3": [], "Col4": []}

based on "COLUMN_NAME" field.

Then you can iterate over the myDF keys (ignoring the "COLUMN_NAME" field ofc) and for each of those keys rule_key you can write something that looks like:

for i, my_dict_key in enumerate(myDF["COLUMN_NAME"]):
    my_dict[my_dict_key].append(myDF[rule_key][i])

Answer 3

d = myDF.set_index('COLUMN_NAME').T.to_dict()

Answer 4

I used your approach to get the dictionary plus some more transformations without null values.

from collections import defaultdict

myDF2 = myDF.T.reset_index(drop=True)

myDF2.columns = myDF2.iloc[0]

myDF3 = myDF2[1:]

dfdict = defaultdict(list)
for (group, col), rule in myDF3.stack().items():
    dfdict[col].append(rule)

dfdict