SQL is dead. The future is MapReduce.Ok, we need SQL on Hadoop. Here's Hive.Hive is too slow. The future is Spark.OK, we need SQL on Spark. Here's Spark SQL.Notice a pattern?Teradata had this figured out in the 1980s. I wrote the full timeline.Read t...

If two companies migrate the same Teradata system to Snowflake, with the same data, indexes, statistics, and workloads, will they get the same bill?They will not. And the reason has less to do with warehouse sizing or SQL translation than most teams...

Your Teradata expertise is valuable. But carried into Snowflake unchanged, it can cost you. This article covers five areas where experienced Teradata professionals get tripped up: your cost model (fixed licensing vs. pay-per-second), your query optim...

I’m excited to introduce DWHPro Query Master – a lightweight, modern web application designed specifically for Teradata users who value speed, simplicity, and powerful features.Soon there will be a demo version available. In the meanwhile p...

Hi,Most people try to fix performance problems by throwing more hardware, more credits, or more magic features at the system.After 20 years tuning some of the largest Teradata and cloud data warehouses in Europe, I’ve learned something much simpler:U...

One of the smallest but most persistent sources of data mismatches during a Teradata-to-Snowflake migration comes from an unassuming function: ROUND(). While both systems follow ANSI SQL semantics, they differ in how they handle values that lie exact...

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 cho...

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 ter...

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...

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 Data...

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...

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 euqals one unit of work. To go fast and cheap, publish many medium parts (about 100–250 MB compressed), size...

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....

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 impo...

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...