转置Python字典的各个元素

Question

I have a python super dictionary that is made of individual dictionaries and currently looks like this:

raw_data1 = {'Series_Date':['2017-03-10','2017-03-13','2017-03-14','2017-03-15'],'SP':[35.6,56.7,41,41],'1M':[-7.8,56,56,-3.4],'3M':[24,-31,53,5]}
raw_data2 = {'Series_Date':['2017-03-10','2017-03-13','2017-03-14','2017-03-15'],'SP':[35.6,56.7,41,41],'1M':[-7.8,56,56,-3.4],'3M':[24,-31,53,5]}
raw_data3 = {'Series_Date':['2017-03-10','2017-03-13','2017-03-14','2017-03-15'],'SP':[35.6,56.7,41,41],'1M':[-7.8,56,56,-3.4],'3M':[24,-31,53,5]}
top_dictionary = {
'raw_data1': raw_data1,
'raw_data2': raw_data2,
'raw_data3': raw_data3
}
print top_dictionary

I would like to transpose the individual dictionaries in my top_dictionary in a way such that all the value fields get transposed to the Value Column and the date is appended as a row item. The column name of the value field becomes a row for the Description column.

As an example the value in top_dictionary with the key: 'raw_data1' will then look like:

raw_data1 = {'Series_Date':['2017-03-10','2017-03-10','2017-03-10','2017-03-13','2017-03-13','2017-03-13','2017-03-14','2017-03-14','2017-03-14','2017-03-15','2017-03-15','2017-03-15'],'Value':[35.6,-7.8,24,56.7,56,-31,41,56,53,41,-3.4,5],'Desc':['SP','1M','3M','SP','1M','3M','SP','1M','3M','SP','1M','3M']}

I know how to do this for each individual dictionary by using the pandas melt function but how do I do this for my top_dictionary such that all elements inside it get transposed accordingly?

Answer 1

Just loop over the top dictionary items. You'll have to play with if you want strict chronological order:

top_dict_new = dict()
for key, data_dict in top_dictionary.items():
    df = pd.melt(pd.DataFrame(data_dict), id_vars = ['Series_Date'])
    top_dict_new[key] = df.to_dict('list')

Edit: This yields:

print top_dict_new['raw_data1']
{'variable': ['1M', '1M', '1M', '1M', '3M', '3M', '3M', '3M', 'SP', 'SP', 'SP', 'SP'], 'Series_Date': ['2017-03-10', '2017-03-13', '2017-03-14', '2017-03-15', '2017-03-10', '2017-03-13', '2017-03-14', '2017-03-15', '2017-03-10', '2017-03-13', '2017-03-14', '2017-03-15'], 'value': [-7.7999999999999998, 56.0, 56.0, -3.3999999999999999, 24.0, -31.0, 53.0, 5.0, 35.600000000000001, 56.700000000000003, 41.0, 41.0]}

Answer 2

Check the "Series_Date" output... if you don't want it sorted, remove the sorted method in the function. I only sorted it to match your specified output.

def make_dict(dict_in, val_cols):
    d_out = {}
    d_out['Value'] = pd.DataFrame(dict_in)[val_cols].values.flatten()
    d_out['Series_Date'] = sorted(dict_in['Series_Date'] * len(val_cols))
    d_out['Desc'] = val_cols * len(val_cols)
    return d_out

new_dict = {k: make_dict(v, ['SP','1M', '3M']) for k, v in top_dictionary.items()}