Create DataFrame from dictionary

Question

I request data from a table in a database and each line comes as a dictionary like this one :

{
 "timestamp" : 1234657890,
 "prices" : {
 "AAA" : 111,
 "BBB" : 222,
 ...
 "ZZZ" : 999
 }
}

From all those lines i wanted to create a dataframe like this:

Timestamp AAA BBB ... ZZZ
1234657890 111 222 ... 999
1234567891 110 223 ... 997
 ...
1324657899 123 208 ... 1024

So i did :

rawData = database_request()
listPrices = []
for row in rawData
 tmp = {'timestamp': row['timestamp']}
 tmp.update({name : price for name,price in row['prices'].items()})
 listPrices.append(tmp)
df = pd.DataFrame(listePrices)

So i was wondering if there were a more pythonic way to do this ?

• user228914: Does it only have prices and timestamps in the dictionary?
• TmSmth: yes only those two fields
• GZ0: You need pd.json_normalize here.

Answer 1

Your rawData (which should be ideally named raw_data, python suggests a style guide to name variables and functions in lower_snake_case) is already in a list structure. You can manipulate this in place, without having to process the whole dataset manually.

for row in raw_data:
 row.update(row.pop("prices"))