KNIME Documentation

Scripting API Reference (knime.scripting.io) ​

The knime.scripting.io module is the primary API for interacting with KNIME data inside Python Script and Python View nodes. It is automatically available on the PYTHONPATH — you do not need to install it.

For the full API reference, see Python Scripting API.

Import ​

python
import knime.scripting.io as knio

Tables ​

SymbolDescription
knio.input_tables[i]Read-only input table at port index i (0-based).
knio.output_tables[i]Assign a knio.Table or knio.BatchOutputTable to write output.
knio.Table.from_pandas(df)Create a KNIME table from a Pandas DataFrame.
knio.Table.from_pyarrow(table)Create a KNIME table from a PyArrow Table.
table.to_pandas()Convert to Pandas DataFrame.
table.to_pyarrow()Convert to PyArrow Table.

Batches ​

SymbolDescription
knio.BatchOutputTable.create()Create an empty batch output table.
input_table.batches()Returns an iterable of batches. Each batch supports .to_pandas() / .to_pyarrow().
batch_output.append(data)Append a Pandas DataFrame or PyArrow Table as a batch.

Objects and images ​

SymbolDescription
knio.input_objects[i]Pickled input object at port index i.
knio.output_objects[i]Assign a picklable object to write output.
knio.output_images[i]Assign an SVG string or PNG bytes to write an image.

Flow variables ​

SymbolDescription
knio.flow_variables['name']Read or write flow variables as a dictionary.

Views (Python View node only) ​

SymbolDescription
knio.output_viewAssign the return value of a knio.view_* function.
knio.view(obj)Auto-detect format and create a view.
knio.view_html(html)View from an HTML string.
knio.view_svg(svg)View from an SVG string.
knio.view_png(data)View from PNG bytes.
knio.view_jpeg(data)View from JPEG bytes.
knio.view_matplotlib(fig?)View from an active or given matplotlib figure.
knio.view_seaborn(fig?)View from an active or given seaborn figure.
knio.view_plotly(fig)View from a Plotly figure. Set custom_data to Row ID for selection sync.
knio.view_ipy_repr(obj)View from an object with an IPython _repr*_ function.

Example: matplotlib ​

python
import knime.scripting.io as knio
import matplotlib.pyplot as plt

fig, ax = plt.subplots()
ax.plot([1, 2, 3], [4, 5, 6])

knio.output_view = knio.view_matplotlib(fig)

Example: Plotly with selection sync ​

To synchronize row selection between a Plotly chart and other KNIME views, set custom_data to the row index:

python
import knime.scripting.io as knio
import plotly.express as px

df = knio.input_tables[0].to_pandas()

fig = px.scatter(
    df,
    x="my_x_col",
    y="my_y_col",
    color="my_label_col",
    custom_data=[df.index],
)

knio.output_view = knio.view_plotly(fig)

Output images from views ​

The optional image output port is populated automatically when the view is SVG, PNG, or JPEG, or can be converted to one:

  • matplotlib / seaborn: PNG or SVG depending on the format argument.
  • Plotly: requires the kaleido package.
  • IPython objects: uses repr_svg, repr_png, or repr_jpeg (first available).
  • HTML: set an image representation manually via the view_html call.

You can also populate knio.output_images[0] directly:

python
import knime.scripting.io as knio
import matplotlib.pyplot as plt
import io

fig, ax = plt.subplots()
ax.bar(["A", "B"], [3, 5])

buf = io.BytesIO()
fig.savefig(buf, format="png")
knio.output_images[0] = buf.getvalue()

Jupyter notebook support ​

SymbolDescription
knime.scripting.jupyter.load_notebook(dir, name)Load a notebook as a Python module.
knime.scripting.jupyter.print_notebook(dir, name)Print cell contents to console.

Next steps ​