Pandas Dataframe to JSON add JSON Object Name - python-3.x

Question

I have a dataframe that I'm converting to JSON but I'm having a hard time naming the object. The code I have:
j = (df_import.groupby(['Item', 'Subinventory', 'TransactionUnitOfMeasure', 'TransactionType', 'TransactionDate', 'TransactionSourceId', 'OrganizationName'])
.apply(lambda x: x[['LotNumber', 'TransactionQuantity']].to_dict('records'))
.reset_index()
.rename(columns={0: 'lotItemLots'})
.to_json(orient='records'))
The result I'm getting:
[
{
"Item": "000400MCI00099",
"OrganizationName": "OR",
"Subinventory": "LAB R",
"TransactionDate": "2021-08-19 00:00:00",
"TransactionSourceId": 3000001595xxxxx,
"TransactionType": "Account Alias Issue",
"TransactionUnitOfMeasure": "EA",
"lotItemLots": [
{
"LotNumber": "00040I",
"TransactionQuantity": -5
}
]
}
]
The result I need (the transactionLines part), but can't figure out:
{
"transactionLines":[
{
"Item":"000400MCI00099",
"Subinventory":"LAB R",
"TransactionQuantity":-5,
"TransactionUnitOfMeasure":"EA",
"TransactionType":"Account Alias Issue",
"TransactionDate":"2021-08-20 00:00:00",
"OrganizationName":"OR",
"TransactionSourceId": 3000001595xxxxx,
"lotItemLots":[{"LotNumber":"00040I", "TransactionQuantity":-5}]
}
]
}
Index,Item Number,5 Digit,Description,Subinventory,Lot Number,Quantity,EOM,[Qty],Transaction Type,Today's Date,Expiration Date,Source Header ID,Lot Interface Number,Transaction Source ID,TransactionType,Organization Name
1,000400MCI00099,40,ACANTHUS OAK LEAF,LAB R,00040I,-5,EA,5,Account Alias Issue,2021/08/25,2002/01/01,160200,160200,3000001595xxxxx,Account Alias Issue,OR
Would appreciate any guidance on how to get the transactionLines name in there. Thank you in advance.

Answer 1

It would seem to me you could simply parse the json output, and then re-form it the way you want:
import pandas as pd
import json
data = [{'itemID': 0, 'itemprice': 100}, {'itemID': 1, 'itemprice': 200}]
data = pd.DataFrame(data)
pd_json = data.to_json(orient='records')
new_collection = [] # store our reformed records
# loop over parsed json, and reshape it the way we want
for record in json.loads(pd_json):
nested = {'transactionLines': [record]} # matching specs of question
new_collection.append(nested)
new_json = json.dumps(new_collection) # convert back to json str
print(new_json)
Which results in:
[
{"transactionLines": [{"itemID": 0, "itemprice": 100}]},
{"transactionLines": [{"itemID": 1, "itemprice": 200}]}
]
Note that of course you could probably do this in a more concise manner, without the intermediate json conversion.