Part 1 of 5 · 1 min read
Why this page comes first
In this module you will
- Define write skew with a concrete, worked scenario.
- Explain why write skew passes every individual transaction's own constraint check, yet still breaks a cross-row invariant.
- Explain how a phantom can still occur under snapshot isolation, despite each transaction's consistent view.
- Recognize the shared root cause behind both anomalies: each transaction only ever sees its own unchanging snapshot.
The previous page closed by predicting that snapshot isolation might not guarantee full serializability. This page delivers on that prediction with two concrete, worked anomalies that snapshot isolation genuinely allows. Each individual transaction still sees a perfectly consistent view of the data.