Streamlit & Snowflake

Building a Streamlit app that displays data from a Snowflake table on screen. My starting point with an example: this Medium post.

The post dives extensively into the Snowflake setup, feel free to read the original post to set this up. Same goes for the setup of the Python venv, the writer does this with Conda, but I do this using the Python venv.

This is the code example, I just did a copy paste from the original post but then decided to change some things, first I swapped the TOML config for a JSON config file.

import streamlit as st 
from snowflake.snowpark import Session 
import json 
 
st.title('❄️ How to connect Streamlit to a Snowflake database') 
 
# connect to Snowflake 
def create_session(): 
    with open('../config/sf-config.json') as f: 
        connection_parameters = json.load(f)   
    session = Session.builder.configs(connection_parameters).create() 
    return session 
 
session = create_session() 
st.success("Connected to Snowflake!")

This is the sf-confg.json layout.

{ 
"account":"", 
"user":"", 
"password":"", 
"warehouse": "", 
"database": "", 
"schema": "" 
}

Start this script using Streamlit Run.

streamlit run sf_connect.py

From this point onwards, you can just put any Streamlit library component in the code to work with the Snowflake tables.

In my case, I added some code to display the table contents in the Streamlit app.

# Load data table 
def load_data(table_name): 
    ## Read in data table 
    st.write(f"Here's some example data from `{table_name}`:") 
    table = session.table(table_name) 
 
    ## Do some computation on it 
    table = table.limit(100) 
     
    ## Collect the results. This will run the query and download the data 
    table = table.collect() 
    return table 
 
# Select and display data table 
table_name = "DB_STREAMLIT.PUBLIC.EMPLOYEE" 
 
## Display data table 
df = load_data(table_name) 
 
## Writing out data 
for row in df: 
    st.write(f"{row[0]}, {row[1]}")

As you can see, 2 records are added.

That’s it for now, more to follow on Streamlit and Snowflake, stay tuned.