KNIME Documentation

Convert data tables ​

Transfer data between KNIME tables and Python structures using the knime.scripting.io (knio) module.

Pandas ​

Pandas is the standard library for data manipulation in Python. Use these methods to work with pd.DataFrame objects.

To convert an input table to a Pandas DataFrame:

python
import knime.scripting.io as knio
df = knio.input_tables[0].to_pandas()

To write a Pandas DataFrame to a KNIME output port:

python
import knime.scripting.io as knio
# Assuming 'df' is your existing Pandas DataFrame
knio.output_tables[0] = knio.Table.from_pandas(df)

PyArrow ​

Apache Arrow provides high-performance, columnar data structures. Use these methods to work with pa.Table objects, which is recommended for larger datasets.

To convert an input table to a PyArrow Table:

python
import knime.scripting.io as knio
table = knio.input_tables[0].to_pyarrow()

To write a PyArrow Table to a KNIME output port:

python
import knime.scripting.io as knio
# Assuming 'table' is your existing PyArrow table
knio.output_tables[0] = knio.Table.from_pyarrow(table)

INFO

The knime package required for these conversions is provided automatically by the default bundled Python environment (see Bundled Packages). Do not install a package named knime via pip in your environment, as this will cause a name clash and script failure.