使用 pandas.io.json.json_normalize 获取“AttributeError: 'float' object has no attribute 'items'”

Question

When trying to normalize a series within a pandas dataframe with the json_normalize function i am getting the Error:

"AttributeError: 'float' object has no attribute 'items'" 

Each row of the series contains a nested json, though some rows don't contain all of the attributes, that are present in some of those json' from that series

also there is a field "timestamp":{"$date":1578411194000} within those nested json's, which is also present in another column of that same dataframe, giving me an error in another attempt to flatten that other series.

I am assuming the AttributeError has something to do with the either not all JSONs containing all the fields or sth. with those timestamps.json_normalize did work for some of the other df-columns.

I hope this is enough info. thanks a lot in advance!

Answer 1

It can happen if there are NaN fields which can be solved by using dropna() :

pd.json_normalize(df.explode("field")["field"]))
=> AttributeError: 'float' object has no attribute 'items'

pd.json_normalize(df.explode("field")["field"].dropna())
=> no error