1. The Forgotten Performance Trick: A Semicolon That Saves Time For decades, Teradata developers have quietly used one of the smallest but most powerful performance optimizations in BTEQ:a semicolon at the start of a line. This isn’t just a style choice.It tells Teradata to combine all statements into one multi-statement request, parsed and executed as …
Performance degradation caused by uneven workload distribution is one of the oldest and most persistent challenges in parallel data warehouse systems. Both Teradata and Snowflake can experience this imbalance, commonly known as skew. Although the term is shared, the mechanics differ fundamentally: Teradata can suffer from both persistent and runtime skew, whereas Snowflake’s skew occurs …
At first glance, an UPDATE looks universal.In reality, it’s one of the most misleading similarities between Teradata and Snowflake.The SQL is the same, but the storage, logging, recovery, and performance mechanics are completely different. If you’re migrating from Teradata to Snowflake (or running both), understanding these differences prevents slow jobs, unnecessary costs, and avoidable outages. …
For many years, Teradata was the undisputed leader in large-scale data warehousing. Banks, insurers, and telcos built their most critical systems on it. Today, the market is very different. Cloud-native databases such as Snowflake, BigQuery, and Databricks have set new standards in elasticity and simplicity. The question is: what role will Teradata play in this …
When migrating analytical workloads from Teradata to Snowflake, one subtle but important performance factor often gets overlooked: how the two systems handle GROUP BY operations on huge tables. The SQL looks the same, but the execution engines behave differently. If you’ve relied on Teradata for years, you may be surprised by Snowflake’s behavior. GROUP BY …
Big files, real risks, and how not to overspend On Snowflake, COPY INTO scales with the number of files, not the total GB. A single big file equals one unit of work. To go fast and cheap, publish many medium parts (about 100–250 MB compressed), size the warehouse to the files in flight, and use …
Snowflake’s physical join execution is predominantly hash-based. In practice you’ll observe hash-join variants with two distributions: If you come from Teradata, the intent will feel familiar: both systems aim to co-locate equal keys before matching. This article explains Snowflake’s strategies, maps them to Teradata’s (including dynamic plan fragments), and shows how to recognize and mitigate …
I was recently approached to support a case of importing the results of a Teradata query into a third-party vendor database. On the export side, Teradata happily wrote close to 60 million rows into a single, wide CSV file. On the other side, the import process to the target database could not cope with the …
Introduction to the Teradata AMP Worker Task The Teradata AMP Worker Task or AWT is the heart of the AMP, responsible for executing tasks and ensuring the smooth functioning of the system. AWTs are threads that process incoming tasks in the AMP. Each AMP has a finite pool of AWTs, which is shared among all …
Introduction to Teradata Performance and NOT NULL Welcome to our latest Teradata performance blog post, a series designed to provide valuable insights into SQL queries. This article spotlights ‘NOT NULL’. To delve deeper into ‘NOT IN’, it is crucial to comprehend a frequently neglected SQL database design principle: properly defining columns that cannot hold NULL …
Expert network for enterprise data platforms. Senior consultants, project teams built for your challenge — across Teradata, Snowflake, Databricks, and more.
📍Vienna, Austria & Jacksonville, Florida
Join 4,000+ data professionals.
Weekly insights on Teradata, Snowflake & data architecture.