If you are getting started with cloud data platforms, one of the first tasks you will face is figuring out how to load data in Snowflake and run queries. Snowflake makes this process straightforward, with options to upload local files, pull from cloud storage, or use connectors such as Salesforce or Google Sheets.
In this guide, we will walk through how to:
- Load a CSV file into Snowflake
- Create a database, schema, and table
- Handle common errors such as invalid characters
- Run SQL queries to analyze results
- Create quick visualizations inside Snowflake
Step 1: Log Into Snowflake and Navigate to Add Data
After logging into your Snowflake account or free trial, go to the Data tab. Select Add Data, which functions similarly to Get Data in Power BI.
Snowflake provides multiple ways to load data:
- Upload a local file
- Load from a stage (temporary storage)
- Use connectors for Google Cloud, Azure, Salesforce, and more
For this example, we will upload a local CSV file.
Step 2: Upload the Data File
We will use failed banks data to answer a real-world question:
Which U.S. bank and state had the most failures?
- Select Load Data into a Table.
- Choose Browse and upload the CSV file.
- Preview the raw data, which contains fields such as bank name, city, state, and closing date.
Step 3: Create a Database, Schema, and Table
To properly store the data, create a structure:
- Database:
failed_banks_DB - Schema: Public (default)
- Table:
failed_banks_table
Snowflake shows a preview of the table structure before loading.
Common issue: If column names contain spaces such as Bank Name, replace them with underscores like bank_name.
Step 4: Define File Format and Error Handling
Snowflake supports formats including CSV, JSON, and Parquet. For this dataset, CSV is sufficient.
Error handling options:
- Load all data, even if errors exist
- Skip invalid rows
- Fail if errors are detected
Step 5: Load the Data
Click Load. Snowflake will create the table and insert the rows. Afterward, you can immediately start querying the data.
Step 6: Run SQL Queries in Snowflake
Snowflake generates a starter query:
SELECT *
FROM failed_banks_DB.public.failed_banks_table
LIMIT 10;
This previews the data. Adjust the LIMIT to retrieve more rows.
Performance is strong. Queries on the full dataset complete in seconds.
Step 7: Analyze Results
With a few SQL statements, we can answer key questions:
- City with the most failed banks: Chicago
- State with the most failed banks: Georgia with 93 failures
Snowflake’s query details also display execution times and performance statistics.
Step 8: Visualize Data in Snowflake
By switching from “Results” to “Chart,” you can quickly visualize your data without leaving the Snowflake interface. This feature is useful for exploratory analysis before exporting to BI tools.
Final Thoughts
In this tutorial, we:
- Created a database, schema, and table
- Uploaded CSV data
- Cleaned column names
- Queried and analyzed results
- Built simple visuals directly in Snowflake
The process to load data in Snowflake is simple, efficient, and scalable. From here, you can integrate with BI tools such as Power BI or Tableau, create transformations, or grant role-based access.
Learn More
Power BI Training by Data Bear
Take your analytics further by combining Snowflake with Power BI.
