PH
Snowflake Dynamic Tables: Declarative Data Pipelines
Published 7/2026
MP4 | Video: h264, 1920x1080 | Audio: AAC, 44.1 KHz, 2 Ch
Language: English | Duration: 8h 31m | Size: 2.92 GB
What you'll learn
Create, configure, and tune Dynamic Table pipelines safely - from first CREATE to chain-level lag design
Master the incremental-refresh support matrix: which joins, aggregations, and window patterns stay incremental and which silently force full refreshes
Design medallion DAGs declaratively - normal-table bronze, chained dynamic-table silver and gold with DOWNSTREAM lag
Implement SCD and history patterns with three proven approaches, including the new CUSTOM_INCREMENTAL refresh mode
Monitor, alert on, and troubleshoot refresh health with the three table functions and native Snowflake alerts
Engineer costs with the three-bucket model: warehouse consolidation, lag tuning, and the traps that silently multiply credits
Reprocess history surgically with the backfill-based rebuild playbook instead of full reinitializations
Use frozen regions and zero-copy backfill to make history immutable and nearly free
Govern dynamic tables with policies, tags, privileges, and lineage across four surfaces
Integrate with dbt, Snowpipe Streaming, and Iceberg - and choose correctly between Dynamic Tables, materialized views, and streams-and-tasks
Deploy changes safely at scale with blue-green swaps, clones, and CI/CD for table definitions
Migrate a real streams-and-tasks pipeline to declarative form, honestly documenting what can and cannot convert
Requirements
A working knowledge of SQL - SELECT, JOIN, GROUP BY, window functions
Basic familiarity with Snowflake (a free 30-day trial covers every exercise in this course - setup instructions included)
Comfortable with the medallion architecture concept (bronze/silver/gold) - not required in advance, but helpful context
No prior experience with Dynamic Tables, streams, or tasks required - we build that from zero
A laptop and a Snowflake account (trial or existing) - no other tools needed
Description
Your streams-and-tasks pipeline breaks every time someone touches the source schema, and you're the only one who understands why. Dynamic Tables fix that - if you know exactly where they help and where they quietly cost you a fortune.
Across 22 modules and 106 lessons you replace fragile, imperative orchestration with declarative pipelines built on Snowflake Dynamic Tables - from your very first CREATE DYNAMIC TABLE to production DAGs with monitoring, alerting, cost control, and blue-green deployment. You follow one data engineer,Priya, who inherits a 200-task DAG nobody can safely change, and rebuild it into something declarative, observable, and honest about its own limits.
What makes this course different
-Every SQL statement is live-verified against a real Snowflake account. Not paraphrased from docs - the exact syntax on screen was tested end-to-end, including a genuinely cutting-edge Public Preview feature (REFRESH_MODE = CUSTOM_INCREMENTAL) most courses don't even know exists yet.
-Honest about the limits, not just the wins. You'll learn exactly which joins, aggregations, and window patterns silently force a full refresh instead of an incremental one - the support matrix most tutorials skip because it's inconvenient.
-Cost engineering as a first-class topic. The three-bucket cost model, the 60-second warehouse-resume floor, and the traps that silently multiply your credit bill - not an afterthought module.
-Production operations, not just CREATE statements. Monitoring, alerting, incident response, schema-evolution restatements, surgical reprocessing, blue-green deployment, and CI/CD - the parts of the job that happen after the demo.
-An honest architect's decision framework. When Dynamic Tables win against materialized views, streams-and-tasks, and dbt - and, just as importantly, when they lose and you should say so.
-A real capstone. Build SNOWMART near-real-time end to end - ingestion, silver, gold, monitoring, and a live failure-injection game day - not a toy exercise.
What you'll build, module by module: your first dynamic table and its two decisions that matter (TARGET_LAG and WAREHOUSE), the refresh-mode landscape (AUTO, FULL, INCREMENTAL, and the CUSTOM_INCREMENTAL preview frontier), the refresh engine's internals, the full incremental-refresh support matrix for joins and aggregations, a declarative medallion DAG, near-real-time ingestion front-ends, SCD and history patterns (three proven approaches, including the new custom-incremental mode), frozen regions and zero-copy backfill, schema evolution and restatements, the surgical reprocessing playbook, monitoring and observability, alerting and incident response, cost engineering, performance tuning and locality, governance and lineage, the ecosystem (dbt, materialized views, DLT, Iceberg), ops at scale (blue-green, clones, failover, CI/CD), an honest migration playbook - and the SNOWMART capstone that ties it all together.
The capstone - SNOWMART Near-Real-Time, End to End: you architect and build a complete medallion pipeline on Dynamic Tables - bronze ingestion, deduped/conformed silver, aggregated gold with an SLA watchtower - then run a blue-green change and a live failure-injection game day, and grade your own work against the same rubric this course teaches.
By the end of this course, you will be able to design, build, operate, and cost-control production Dynamic Table pipelines - and know exactly when NOT to use them.
Enrol now. Declarative pipelines are the future of Snowflake data engineering - build the judgment to use them correctly, not just the syntax to turn them on.
Who this course is for
Data engineers maintaining fragile streams-and-tasks pipelines who want a declarative alternative
Analytics engineers and Snowflake practitioners who want to move beyond basic CREATE TABLE AS
Data platform engineers who need to make an honest build-vs-Dynamic-Tables call for their team
Anyone preparing to operate Dynamic Tables in production - not just demo them
SnowPro-track engineers who want hands-on depth beyond exam-surface knowledge