Streamlit

Streamlit#

To display Python DataFrames as interactive DataTables in Streamlit applications, use

from itables.streamlit import interactive_table

We have a sample application available at https://itables.streamlit.app (source code here)

Selected rows#

This feature was added in ITables v2.2.0.

Use the selected_rows: list[int] argument from interactive_table to select rows when the table is first displayed. Add select=True to let the user modify the selection. Then, the interactive_table component returns a dict, with a key "selected_rows" that points to the updated selection.

Limitations#

In most cases, you will be able to use interactive_table in a Streamlit application in the same way that you use show in notebooks.

Due to implementation constraints, the Streamlit component has some limitations that show does not have:

  • Pandas Style objects can’t be rendered with interactive_table. This is because the Streamlit component needs to pass the table data to the frontend in JSON format (while Pandas Style objects are formatted using HTML)

  • Similarly, you can’t use the use_to_html argument in interactive_table

  • Complex column headers might look different than in notebooks, and HTML in columns is not supported

  • JavaScript callbacks like custom formatting functions are not supported

  • The interactive table is rendered within an iframe that has a fixed weight. This does not work well with the lengthMenu control, nor with the advanced filtering extensions (if that is an issue for you, please subscribe or contribute to #275).

Workaround#

If you hit one of the limitations above, you can fallback to using to_html_datatable in combination with Streamlit’s html function.

Please note that:

  • you will have to specify the table height manually,

  • an internet connection is required when using to_html_datatable,

  • the app/table might take longer to display.

A sample application is available at https://to-html-datatable.streamlit.app (source code here)

Contents