将嵌套字典转换为数据框

Question

My dictionary looks like this

mydict = 
{240594.0: {1322.0: 1.6899999999999999, 1323.0: 1.6900000000000002, 1324.0: 1.6899999999999999, 1325.0: 1.6899999999999999, 1326.0: 1.6899999999999999, 1327.0: 1.6900000000000002, 1328.0: 1.6899999999999999, 1329.0: 1.6899999999999999, 1356.0: 1.6900000000000002, 1357.0: 1.6900000000000002, 1358.0: 1.6899999999999999, 1359.0: 1.6900000000000002, 1360.0: 1.6900000000000002, ...},

226918.0: {1322.0: 1.6900000000000002, 1323.0: 1.6899999999999999, 1324.0: 1.6900000000000002, 1325.0: 1.6899999999999999, 1326.0: 1.6900000000000002, 1327.0: 1.6899999999999999, 1328.0: 1.6900000000000002, 1329.0: 1.6899999999999999, 1352.0: 1.6900000000000002, 1353.0: 1.6900000000000002, 1354.0: 1.6899999999999999 ...}}

which is the real value of {iri_key: {week:price, week:price ...}, iri_key: {...}} and I want to convert this dictionary into dataframe which looks like

         week week  week ...
irikey: price price price ...
irikey: ...    ...   ...

in above case

           1322.0                  ...
240594.0   1.6899999999999999      ...
226918.0   1.6900000000000002      ...

how could I do this?

Answer 1

As you have probably discovered, DataFrame(mydict) is valid code. You could simply take the transpose ( .T ) to get your desired result.

A better way, in terms of code readability and directness, is available: use the specific DataFrame constructor DataFrame.from_dict , which has a keyword argument orient .

In [2]: DataFrame.from_dict(mydict, orient='index')
Out[2]: 
        1356  1357  1358  1359  1360  1322  1323  1324  1325  1326  1327  \
226918   NaN   NaN   NaN   NaN   NaN  1.69  1.69  1.69  1.69  1.69  1.69   
240594  1.69  1.69  1.69  1.69  1.69  1.69  1.69  1.69  1.69  1.69  1.69   

        1328  1329  1352  1353  1354  
226918  1.69  1.69  1.69  1.69  1.69  
240594  1.69  1.69   NaN   NaN   NaN  

[2 rows x 16 columns]

As you can see from the example data you provided, missing values and variable lengths are handled properly.