DataFrame#

Open this notebook in Jupyterlite | Download this notebook from GitHub (right-click to download).

import pandas as pd
import panel as pn

pn.extension()

The DataFrame pane renders pandas, dask and streamz DataFrame and Series types as an HTML table. Other DataFrame-like objects supported by Narwhals, e.g. polars and pyarrow, are also supported by converting them to pandas first. The Pane supports all the arguments to the DataFrame.to_html function.


If you need to to display a larger DataFrame or use advanced table features and interactivity we recommend using the Tabulator widget or Perspective pane instead.




Parameters:#


For details on other options for customizing the component see the layout and styling how-to guides.


    

  • bold_rows (boolean, default=True): Make the row labels bold in the output.
    • 

  • border (int, default=0): Width of the border included in the opening table tag.
    • 

  • classes (list[str]): CSS class(es) to apply to the resulting html table.
    • 

  • col_space (int, str or dict): The minimum width of each column in CSS length units. An int is assumed to be px units.
    • 

  • decimal (str, default=’.’): Character recognized as decimal separator, e.g. ‘,’ in Europe.
    • 

  • escape (boolean, default=True): Convert the characters <, >, and & to HTML-safe sequences.
    • 

  • float_format (function): Formatter function to apply to columns’ elements if they are floats. The result of this function must be a unicode string.
    • 

  • formatters (dict or list): Formatter functions to apply to columns’ elements by position or name. The result of each function must be a unicode string.
    • 

  • object (object): The DataFrame object being displayed
    • 

  • header (boolean, default=True): Whether to print column labels.
    • 

  • index (boolean, default=True): Whether to print index (row) labels.
    • 

  • index_names (boolean, default=True): Whether to print the names of the indexes.
    • 

  • justify (str): How to justify the column labels (‘left’, ‘right’, ‘center’, ‘justify’, ‘justify-all’, ‘start’, ‘end’, ‘inherit’, ‘match-parent’, ‘initial’, ‘unset’)
    • 

  • max_rows (int): Maximum number of rows to display.
    • 

  • max_cols (int): Maximum number of columns to display.
    • 

  • na_rep (str, default=’NaN’): String representation of NAN to use.
    • 

  • render_links (boolean, default=False): Convert URLs to HTML links.
    • 

  • show_dimensions (boolean, default=False): Display DataFrame dimensions (number of rows by number of columns).
    • 

  • sparsify (boolean, default=True): Set to False for a DataFrame with a hierarchical index to print every multi-index key at each row.
    • 

  • text_align (str): How to justify the non-header cells (‘left’, ‘right’, ‘center’)
    • 





The DataFrame uses the inbuilt HTML repr to render the underlying DataFrame:


df = pd.DataFrame({
    'int': [1, 2, 3],
    'float': [3.14, 6.28, 9.42],
    'str': ['A', 'B', 'C'],
    'bool': [True, False, True],
}, index=[1, 2, 3])

df_pane = pn.pane.DataFrame(df, width=400)

df_pane







            
            
Like all other Panel objects changing a parameter will update the view allowing us to control the styling of the dataframe. In this example we will control these parameters using a set of widgets created directly from Pane:


pn.panel(df_pane.param, parameters=['bold_rows', 'index', 'header', 'max_rows', 'show_dimensions'],
         widgets={'max_rows': {'start': 1, 'end': len(df), 'value': len(df)}})







            
            
By setting escape to False you can include HTML markup in your Dataframe pane


links = pd.DataFrame({
    "site": ["Docs", "Discourse", "Github", "Twitter"], 
    "url": ["https://panel.holoviz.org/", "https://discourse.holoviz.org/", "https://github.com/holoviz/panel", "https://twitter.com/Panel_org"]
})
links["value"]="<a href='" + links["url"] + "' target='_blank'>" + links["site"] + "</a>"
pn.pane.DataFrame(links, escape=False, width=800, index=False)







            
            




Larger DataFrames#


For larger dataframes it can be useful to set sizing_mode="stretch_both" to make sure they don’t overflow. When you do this, you can use max_height to specify the (maximum) height.


from bokeh.sampledata.airport_routes import airports





table = pn.pane.DataFrame(airports.head(50), sizing_mode="stretch_both", max_height=300)
table







            
            
pn.Column("## Header", table, "## Footer", height=400, width=500).servable()







            
            




Streamz DataFrames#


In addition to rendering standard pandas DataFrame and Series types the DataFrame pane will also render updating streamz types.
Note:


    

  • In a live kernel you should see the dataframe update every 0.5 seconds.
    • 

  • streamz does currently not work in Pyodide/PyScript.
    • 



from streamz.dataframe import Random

sdf = Random(interval='200ms', freq='50ms')

pn.pane.DataFrame(sdf, width=500)







            
            
type(sdf.groupby('y').sum().x)





<class 'streamz.dataframe.core.Seriess'>




Open this notebook in Jupyterlite | Download this notebook from GitHub (right-click to download).