from_pandas(data_frame) conversion failed

Hello,

I’m developing a node in Python reading from a text file, after creating the pandas dataframe, when calling the “knext.Table.from_pandas(df_full)” it returns me an error:

(‘Could not convert 20.0 with type float: tried to convert to boolean’, ‘Conversion failed for column Base with type object’)

any idea?

Hi,

cool that you give Python node development a try!

Could you provide a minimal example in GitHub and a minimal workflow?

Or show the types and content of the pandas df df_full?

Best regards
Steffen

2 Likes

Hello Steffen, thanks for your quick answer,

we’re much more skilled at python compared to Java, so we’re giving it a go!

regarding the problem here an extraction of the dataframe, problem is the dataframes always varies,
I’ve solved the problem converting ALL to string, but it’s strange that the .from_pandas() doesn’t handle potential conversion errors

out.csv (41.5 KB)

Hi @emilianocapasso,

I don’t understand: if I use your file and have the following in my execute method, then I get the screenshot as expected. I used KNIME Analytics Platform (KAP) 4.7.4. Could you elaborate how this failed on your machine and also which KAP and OS you use?

        df = pd.read_csv("<path/to/my>/out.csv")
        return knext.Table.from_pandas(df)

Thanks!

Best regards
Steffen

1 Like

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.