How to convert Dataframe into key value pair list for every single raw?

What is the best way to convert following pandas dataframe to a key value pair like below example:

dataframe:

   col1   col2  col3         col4
0    a    tom    90         Good
1    b    jay    75  Not so good
2    c  harsh    35          Bad
3    d    ras    50    impresive

it is possible to get lists like this
desired ouput:

[
 {
   'key':'col1',
   'value':'a'
 }
{
   'key':'col2',
   'value':'tom'
 }
{
   'key':'col3',
   'value':90,
 }
{
   'key':'col4',
   'value':'Good',
 }
]
[
 {
   'key':'col1',
   'value':'b'
 }
{
   'key':'col2',
   'value':'jay'
 }
{
   'key':'col3',
   'value':75,
 }
{
   'key':'col4',
   'value':'Not so good',
 }
]

…
for every raw

Thank you in advance

>Solution :

A more pythonic way:

>>> [[{'key': k, 'value': v} for v in l]
        for _, k, l in zip(*df.to_dict('split').values())]

[[{'key': 'col1', 'value': 'a'},
  {'key': 'col1', 'value': 'tom'},
  {'key': 'col1', 'value': 90},
  {'key': 'col1', 'value': 'Good'}],
 [{'key': 'col2', 'value': 'b'},
  {'key': 'col2', 'value': 'jay'},
  {'key': 'col2', 'value': 75},
  {'key': 'col2', 'value': 'Not so good'}],
 [{'key': 'col3', 'value': 'c'},
  {'key': 'col3', 'value': 'harsh'},
  {'key': 'col3', 'value': 35},
  {'key': 'col3', 'value': 'Bad'}],
 [{'key': 'col4', 'value': 'd'},
  {'key': 'col4', 'value': 'ras'},
  {'key': 'col4', 'value': 50},
  {'key': 'col4', 'value': 'impresive'}]]

Old answer
Use melt reshape your dataframe then group by index and finally export each group to a list of dict:

>>> [df.to_dict('records') for _, df in df.melt(var_name='key', value_name='value', ignore_index=False).groupby(level=0)]

[[{'key': 'col1', 'value': 'a'},
  {'key': 'col2', 'value': 'tom'},
  {'key': 'col3', 'value': 90},
  {'key': 'col4', 'value': 'Good'}],
 [{'key': 'col1', 'value': 'b'},
  {'key': 'col2', 'value': 'jay'},
  {'key': 'col3', 'value': 75},
  {'key': 'col4', 'value': 'Not so good'}],
 [{'key': 'col1', 'value': 'c'},
  {'key': 'col2', 'value': 'harsh'},
  {'key': 'col3', 'value': 35},
  {'key': 'col4', 'value': 'Bad'}],
 [{'key': 'col1', 'value': 'd'},
  {'key': 'col2', 'value': 'ras'},
  {'key': 'col3', 'value': 50},
  {'key': 'col4', 'value': 'impresive'}]]