DataFrameField Quirks

The greatest quirk in the DataFrameField class comes from the fact that the Pandas DataFrame instances are coerced to a python dictionary via df.to_dict("list") as soon as it is bound to a MongoEngine.Document instance, and likewise a new Pandas DataFram instance is created from the stored dict when it is retrieved via dot notation. As result, when it is retrieved from document class via dot notation, a different instance is returned.

For example:

import pandas as pd
from mongoengine import Document
class my_doc(Document)
    df = DataFrameField()

# CREATE THE INVENTORY REQUEST DataFrame
df1 = pd.DataFrame({
    'goods': ['a', 'a', 'b', 'b', 'b'],
    'stock': [5, 10, 30, 40, 10],
    'category': ['c1', 'c2', 'c1', 'c2', 'c1'],
    'date': pd.to_datetime(['2014-01-01', '2014-02-01', '2014-01-06', '2014-02-09', '2014-03-09'])
})

doc_inst = my_doc()

doc_inst.df = df1
df2 = doc_inst.df

print(df1 is df2)
#> False -- df1 and df2 are two separate Pandas DataFrame instances

jdthorpe/DataFrameField-quirks.md

DataFrameField Quirks