Search code examples
pythonexasol

Pandas timestamps in ISO format cause Exasol error when importing

When using pyexasol's import_from_pandas(df) for a DataFrame, df, which has a datetime column, Exasol (6.2) throws an error because it can't parse the ISO-formatted string representation of the dataframe column. Specifically, the "+00:00" final characters are unparsable by Exasol. My current workaround is to turn all pandas datetime columns into string columns, but that can cost a lot of time.

What's the right way to import datetime columns from Pandas dataframes into an existing Exasol table with a TIMESTAMP column type?

Solution

  • PyEXASOL creator is here.

    You may use import_params dictionary argument to pass additional parameters to pandas.to_csv() method which is used internally. One of such parameters is date_format. Just pass the right format compatible with Exasol.

    I'll consider adding this parameter by default.

    Hope it helps!