Table (or grid) with multiple inputs

🧩 Custom Components
1

Hi,

I’m new to Streamlit and it’s awesome so far. However, I feel I’ve reached the limit of what I am capable of without asking for help. My requirement is a simple one, but it seems it falls outside of the capabilities of the components and how they work currenlty. I feel I’m asking that bit too much…

Basically, I have some contacts in a database, and I want to record whether they have either paid and/or attended an event by adding two columns into a table (or grid) as shown in the example below. I already have the data in a table, no issues, it’s just adding the controls in to two columns at the end of the table that I need. And possibly how to update records in the database, although I’m confident that once I get over this initial hurdle I’ll be able to figure that bit out.

I’ve looked at columns, tables and aggrid so far, but it’s getting over my head. Am I missing something simple here, or should I move on to something that’s not Streamlit that would better suit my needs?

table
table1610×299 86.5 KB

Thanks for reading!

1 Like
2

There is a sample code to add checkboxes in the column.

I tried experimenting it and it worked.

streamlit==1.17.0
streamlit-aggrid==0.3.3

import streamlit as st
import pandas as pd
from st_aggrid import AgGrid, GridOptionsBuilder, JsCode


data = {
    'Name': ['a', 'b', 'c'],
    'Paid': [True, False, True],
    'Attended': [False, True, True]
}

checkbox_renderer = JsCode("""
class CheckboxRenderer{

    init(params) {
        this.params = params;

        this.eGui = document.createElement('input');
        this.eGui.type = 'checkbox';
        this.eGui.checked = params.value;

        this.checkedHandler = this.checkedHandler.bind(this);
        this.eGui.addEventListener('click', this.checkedHandler);
    }

    checkedHandler(e) {
        let checked = e.target.checked;
        let colId = this.params.column.colId;
        this.params.node.setDataValue(colId, checked);
    }

    getGui(params) {
        return this.eGui;
    }

    destroy(params) {
    this.eGui.removeEventListener('click', this.checkedHandler);
    }
}//end class
""")

df = pd.DataFrame(data)

st.write('#### init data')
st.dataframe(df)

gb = GridOptionsBuilder.from_dataframe(df)
gb.configure_column('Paid', editable=True, cellRenderer=checkbox_renderer)
gb.configure_column('Attended', editable=True, cellRenderer=checkbox_renderer)

st.write('#### interface')
ag = AgGrid(
    df, 
    gridOptions=gb.build(),
    allow_unsafe_jscode=True,
    enable_enterprise_modules=False
)

new_data = ag['data']

st.write('#### updated data')
st.dataframe(new_data)

You can do whatever you want with the new_data.

Output

image
image620×685 14.6 KB

3 Likes
3

This is brilliant! Exactly what I needed, thank you very much. :smiling_face_with_three_hearts:

For saving the resulting changes to the database, would I iterate throguh the data in new_data and write each change to the database, or is there a more efficient way? One thing I’m noticing is that for every 10 lines of code in .Net (I’m used to .Net development), there is a single line in Python!! Old habits die hard.

Thanks for your help thus far!

4

We have init and updated dataframe, we can get the difference and that would reduce your iteration.

Add the following code below the existing code.

st.write('#### Changed')
diff = new_data[~new_data.apply(tuple,1).isin(df.apply(tuple,1))]
st.dataframe(diff)

You may iterate on the diff.

Output

image
image506×854 17.4 KB

5

Hey, thanks again. That works, but it’s adding lots of duplicate entires into the database. I think I’m missing a fundamental understanding of how Streamlit works at this point. I just want to commit the entries in the upadated data (diff) dataframe and reload the page, yet the dataframe keeps getting bigger the more changes I add.

6

What database do you use?

7

Hi! I’m using Microsoft SQL Server. I have made great progress on this over the weekend, and I just wanted to say thank you again for your help.

1 Like
8

That’s working swimmingly, nice.