Top 10 Ultimate Challenges in Snowflake Hands-On

Use one sql statement to create the fruits table with ctas, then copy 'apple' into apples and 'orange' into oranges, yielding apples with two entries and oranges with one.

Demonstrate multi-table inserts in Snowflake by selecting names from fruits and routing them to apples or oranges with a conditional insert, plus overwrite and else coverage across Fruits5 and FruitsElse.

Uses a single sql statement with Snowflake's multi-table insert, demonstrating insert first for apples, orange entries into oranges, insert all to copy into apples and fruits5, and an else.

Extend the pivot query by adding a gender column with male and female values, creating headers: country, married, male, married female, single male, and single female. Experiment with a second pivoted column to produce four total columns, then compare solutions after testing.

Explore Snowflake pivot queries and their limit to a single pivoted column and a single aggregated value, then emulate pivots with group by and case when.

Examine the pivot limitations in Snowflake, and learn to emulate multiple pivot and aggregate columns with a group by, using a CTE and hardcoded headers.

Discover how to determine if Jack and Mary like the same food without revealing each preference, using Snowflake vault tables and a secure, private comparison.

Explore how to implement data clean rooms in Snowflake by using a row access policy to compare secrets between Jack and Mary without exposing them.

Explore Yao's millionaires problem and how a data clean room enables private collaboration to compute aggregates like shared customers without revealing sensitive data.

Discover the easiest way to add, delete, or modify table data in snowsight without sql, with the option to type python and hands-on experimentation to test solutions.

Show how to build an in-place data editor for a Snowflake customers table using snowsight, streamlit, and snowpark, transferring edits from a pandas data frame back to Snowflake with write_pandas.

Compare data warehouses to oltp databases, noting ingest via large data files and the read-only nature of Snowsight. Learn how Streamlit apps extend Snowflake with easy data editing.

Learn how to display in Snowsight a graph of the tables used by every view in your Snowflake account, and navigate common visibility issues when the graph isn't easily found.

Copy and run setup script to explore Snowflake object dependencies, focusing on views referencing tables, including materialized views and fully qualified names, and render the graph with Streamlit using Graphviz.

Explore hierarchical metadata in Snowflake, including role hierarchies, task graphs, query profiles, and data lineage, with quick prototyping via Streamlit and Snowpark for immediate visualizations.

Create a customers table with name and country text columns, insert data for ten minutes, display number of entries added each minute, then explore how Snowflake audits these operations.

Explore how to generate and analyze insert counts over time in Snowflake using time travel, lag, and the information_schema query history to create a minute-by-minute distribution chart.

Time travel serves data recovery with limitations. Snowflake permanently audits every SQL operation; inspect real-time query history in information schema and use account usage for longer history despite latency.

Build a snowflake challenge: a parent-child hierarchy from an employees table with employee and manager columns, listing each name indented under its manager, with the president as the top node.

Explore Snowflake's hierarchical queries, including sys_connect_by_path and recursive CTEs, to build employee hierarchies with path and indentation.

discover how to render hierarchies in Snowflake using recursive CTEs and recursive views, compare legacy connect by and level functions, and leverage multi-level joins with depth awareness.

Generate one million rows of realistic customer data in Snowflake, explore scalable data generation, label each column by data type, and compare generator and random functions with data governance considerations.

Learn to generate synthetic data with Faker, create and save a 1 million-row dataset in pandas, and label Snowflake columns with semantic categories via Snowpark and Python worksheets.

Generate fake or synthetic data to address privacy and data science needs, demonstrate data augmentation, and avoid real personal details using Faker, a pandas data frame, and Streamlit in Snowflake.

Compare user-managed tasks with a dedicated warehouse to serverless tasks, choosing the Snowflake implementation for quick tasks: one per minute and one every two minutes, while considering compute costs.

Explore hands-on Snowflake task scheduling and cost analysis by creating user-managed and serverless tasks (T1–T4), monitoring execution times, warehouse metering, and credits in a free test account.

Learn how to optimize Snowflake compute costs by managing virtual warehouses, choosing between user-managed and serverless tasks, and understanding billing by the second.