Convert this list of dictionaries into a list of panda dataframes based on key-value pair in list of-CodePudding

I have this list of dictionaries in python.

l=[
    {
        "TIME": "20220414 12:00:00",
        "Ticker": "AAA",
        "value": "12.0"
    },
    {
        "TIME": "20220414 11:00:00",
        "Ticker": "AAA",
        "value": "13.0"
    },
    {
        "TIME": "20220414 12:00:00",
        "Ticker": "BBB",
        "value": "15.0"
    },
    {
        "TIME": "20220414 11:00:00",
        "Ticker": "BBB",
        "value": "16.0"
    },
    {
        "TIME": "20220414 12:00:00",
        "Ticker": "CCC",
        "value": "17.0"
    },
    {
        "TIME": "20220414 11:00:00",
        "Ticker": "CCC",
        "value": "18.0"
    }   
]

I want to convert it into a list of panda dataframes based on Ticker.

The converted output will look something like this;

The list of 3 dictionaries will be converted into a list of 3 panda dataframes.

I am using python 3.9

CodePudding user response：

You can use the DataFrame constructor combined with groupby and a list comprehension:

out = [d.set_index('Ticker') for _,d in pd.DataFrame(l).groupby('Ticker')]

output:

[                     TIME value
 Ticker                         
 AAA     20220414 12:00:00  12.0
 AAA     20220414 11:00:00  13.0,
                      TIME value
 Ticker                         
 BBB     20220414 12:00:00  15.0
 BBB     20220414 11:00:00  16.0,
                      TIME value
 Ticker                         
 CCC     20220414 12:00:00  17.0
 CCC     20220414 11:00:00  18.0]

NB. I assumed "Name" in the first dictionary should be "Ticker" (and fixed the typo in your question). If this was the real data and not a typo, you first need to change "Name" into "Ticker":

out = [d.set_index('Ticker') for _,d in
       pd.DataFrame(l).assign(Ticker=lambda d: d['Ticker'].fillna(d.pop('Name')))]