Hi, everyone
I'm trying to load sequences using Python SDK by inserting dataframes. In the dataframe, I have some columns with NaN and numeric values.
Dataframe:
![](https://uploads-eu-west-1.insided.com/cognite-en/attachment/4c88af10-2386-4ea4-8a87-9d11dfb7b5bf.png)
and when loading NaN values, they are uploaded as 'null'. I've tried replacing them with empty strings, but due to the column definitions, if I have numeric values, the column's data type must be Double or Long, and conflicts arise when encountering empty strings.
This is the code to load a sequence:
![](https://uploads-eu-west-1.insided.com/cognite-en/attachment/ea3247d3-f018-42b1-980d-a0ad45057efb.png)
Results in CDF using Python SDK:
![](https://uploads-eu-west-1.insided.com/cognite-en/attachment/7ea7dd25-a7e4-420e-949e-80ed13d06213.png)
However, when loading the sequence via DB-Extractor, empty strings are uploaded as such and not as 'null'
![](https://uploads-eu-west-1.insided.com/cognite-en/attachment/5f8be269-931c-4e24-9c43-da5852b2be3a.png)
Is there any way to load via SDK so that NaN values are uploaded as empty and not 'null'?
Regards,
Karina Saylema
Best answer by Jason Dressel
View original