Using style to color rows impact in numbers format in my st.dataframe

Using Streamlit
1

Morning, guys.

I’ve recently colored the rows of my st.dataframe using .style. But all columns with numbers data changed from 6.42, e.g., to 6.42000000. Only columns with string format remained unchanged.
Is there a way to apply .style without changing the original data columns format?

Code used:

prev_data = None
row_styles = []
current_style = None

if len(df6_design['Gama'].unique()) > 1:
    # Loop through rows and apply row styles
    for index, row in df6_design.iterrows():
        current_data = row['Gama']

        if current_data != prev_data:
            # Change the current style when the product code changes
            current_style = 'background-color: #ffffff' if current_style != 'background-color: #ffffff' else 'background-color: #f2f7f2'

        row_styles.append(current_style)
        prev_data = current_data
else:
    # Loop through rows and apply row styles
    for index, row in df6_design.iterrows():
        current_data = row['Descrição']

        if current_data != prev_data:
            # Change the current style when the product code changes
            current_style = 'background-color: #ffffff' if current_style != 'background-color: #ffffff' else 'background-color: #f2f7f2'

        row_styles.append(current_style)
        prev_data = current_data
 
df7_final = df6_design.style.apply(lambda x: row_styles, axis=0)
st.dataframe(df7_final, use_container_width=True, hide_index=True)
2

Hi @Joao_Couto

You can use the format parameter in st.column_config.NumberColumn to format the columns of st.dataframe. For example, valid formatters include %d %e %f %g %i %u.

Here is the link to the Docs page with further information and example:

Hope this helps!

3

Hey @Joao_Couto I had the same issue, posted here a few days ago. Not just that highlighting changes the number format (which I also observed), but that column_config formatter then doesn’t work when the input is df.style.

4

Thanks for sharing, @lawtj. The solution i got to work was using:

df6_design['Jan-Set (Qtd)'] = df6_design['Jan-Set (Qtd)'].apply(
    '{:,.0f}'.format)

This forces the column to be in a string format to show the number as 6.42 e not 6.420000000. But, unfortunately, i just need this column to keep its float format, because, if I want to interact with the dataframe, trying to sort this column, for example, it will understand that 9.33 is higher than 81.40, because 9 is higher than 8 in alphabetical sorting.

I will try what dataprofessor said above and see if it works. I’ll let you know if anything changes.

5

Hello, @dataprofessor. Thanks for your time.

I tried to implement in my last code line this update:

st.dataframe(df7_final, column_config={
        'Jan-Set (Qtd)': st.column_config.NumberColumn(
            'Quantity',
            format='%d'
        )
    }, use_container_width=True, hide_index=True, 
)

The title name changed to ‘Quantity’, but the number format remained 6.420000.

Then i tried to change to:

st.data_editor(df7_final, column_config={
        'Jan-Set (Qtd)': st.column_config.NumberColumn(
            'Quantity',
            format='%d'
        )
    }, use_container_width=True, hide_index=True, 
)

And this worked, the numbers came back to its proper format. BUT, the style thing i did to color the rows disappeared. And i have to say that data_editor style with this bars and pencil in the header is not my thing, i wish i could remain the color style thing and number format using st.dataframe, not data_editor.

The solution i got to use was explained to lawtj in the answers below. But, like I said, it’s not the solution i wanted, because it prohibits me from using the sorting thing of a dataframe.

Thanks again for your time.

6

Hi @Joao_Couto

In that case, could you round the numerical values using Pandas’s built-in function round(), please see this excellent article on how:

Hope this helps!

7

Hi, @dataprofessor . Thanks for your time.

This happens when i use Round(): TypeError: ‘Styler’ object is not subscriptable

And that’s because Styler does not return a dataframe, but a style object. From there, I started to use apply functions with format(). But, unfortunately it changes to a string format.

8

I also ran into this issue, am finding solution online now.
Have u solved it yet?

9

Hello, @aiya6502 . Thanks for commenting. I haven’t solved it. Not the way I wanted. The solution I got was the one I told @lawtj in my answer above.

10

Hi, everyone (@aiya6502, @dataprofessor, @lawtj)!

I’ve got the best solution to solve all my issues mentioned above.

This snippet below adjusts the format even after applying ‘.style’.
Check if it’s gonna help you out.

df7_final = df6_design.style.apply(lambda x: row_styles, axis=0)

df7_final = df7_final.format({'% N-1 (Qtd)': "{:,.1f}".format,
                                '% N-1 (€)': "{:,.1f}".format,
                                'Jan-Set (Qtd)': "{:,.0f}".format,
                                'Prev. Out-Dez (Qtd)': "{:,.0f}".format,
                                'Total Ano (Qtd)': "{:,.0f}".format})

st.dataframe(df7_final, use_container_width=True, hide_index=True)
1 Like
11

This solution work for me, Thank you Joao_Couto

1 Like
12

This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.