Add an on_complete or on_focus_leave for st.data_editor to debounce saving state

[Ben-Epstein]

Checklist

• I have searched the existing issues for similar feature requests.
• I added a descriptive title and summary to this issue.

Summary

I'm looking for a trigger for data_editor that fires only when the user leaves focus of the table, rather than on every change to every cell. Kind of like a debounce to the table.

Why?

I want to use the data editor in a multi page application and keep it persisted across pages in an intuitive way.

The naive thought would be

st.session_state["my_data"] = st.data_editor(st.session_state["my_data"], ...)

You cant do this because it causes the table to reset whenever you change a value. If I add a row, it then gets removed when I move to a new cell/click out.

There is a key and on_change but that has the same issue, wherein I can do something like

def _update_state():
    base_df = se.session_state["my_data"].copy()
    edited_rows = st.session_state.get("ed", {})
    added = edited_rows.get("added_rows", [])
    edited = edited_rows.get("edited_rows", {})
    deleted_idxs = edited_rows.get("deleted_rows", [])
    if deleted_idxs:
        base_df = base_df.drop(deleted_idxs)
    for idx, changes in edited.items():
        for col, changed_value in changes.items():
            base_df.loc[idx, col] = changed_value
    for new_row in added:
        ... # add new rows
    if not st.session_state["my_data"].equals(base_df):
        st.session_state["my_data"] = base_df

st.data_editor(st.session_state["my_data"], key="ed", on_change=_update_state)

but when you update my_data in session state, it breaks out the users focus from the data editing table.

How?

What i'd love is a trigger on_complete or on_focus_remove where the user actively chooses to click out of the table, then the trigger fires. In that world, my _update_state would actually work perfectly.

Additional Context

This relates to #7749 but i wanted to formalize the request

[github-actions]

To help Streamlit prioritize this feature, react with a 👍 (thumbs up emoji) to the initial post.

Your vote helps us identify which enhancements matter most to our users.

[amanchaudhary-95]

I've the same issue. Every time I try to update the dataframe using the on_change callback, it breaks out the focus from the data editing table. It's not ideal from UI perspective.

[Ben-Epstein]

This is actually how the behavior of text_area works today. on_change only takes effect when the user clicks out of scope of the text area box.