What Is Apache Iceberg?
Apache Iceberg is an open table format for very large analytic datasets. It sits on top of files stored in a data lake, Parquet files on cloud storage, for example, and adds the structure and guarantees that make those files behave like a real database table. With Iceberg, data stored cheaply in open files gains reliable transactions, schema changes, and consistent reads, which is much of what turns a data lake into a lakehouse.
The Problem Iceberg Solves
Storing data as files in a data lake is cheap and open, but raw files lack the guarantees a database provides. Without a table format, concurrent writes can corrupt data, schema changes are risky, and there is no clean way to see a consistent snapshot. Iceberg adds a metadata layer over the files that tracks which files make up a table at any moment, so the data behaves transactionally even though it is just files underneath.
Key Features
Iceberg is known for a few capabilities in particular:
- ACID transactions, so concurrent reads and writes stay consistent.
- Schema evolution, so columns can be added, dropped, or renamed without rewriting the data.
- Hidden partitioning, so queries are fast without users having to know the physical layout.
- Time travel, so a query can read the table as it existed at an earlier point.
Together these give file-based data the reliability that used to require a traditional warehouse.
Iceberg and the Lakehouse
Open table formats like Iceberg are a core part of what makes the lakehouse possible. The lakehouse promises warehouse reliability on open, low-cost lake storage, and the table format is the piece that delivers the reliability. Iceberg is one of the leading formats, alongside others such as Delta Lake, and a growing number of engines and platforms can read and write it, which is part of its appeal: data in Iceberg is not locked to one vendor’s engine.
Iceberg and Processing Engines
A table format and a processing engine work together. Iceberg defines how the table is structured and tracked; an engine such as Apache Spark reads and writes the data according to that definition. Because Iceberg is open, many engines can operate on the same tables, so an organization can choose its processing tools without copying the data into a proprietary format first.
Open Formats in an ERP Data Foundation
For companies consolidating ERP data, an open table format means the foundation is not tied to a single vendor. Data landed and modeled once can be read by the tools the business chooses, now and later. QuickLaunch builds governed foundations on open lakehouse architecture for JD Edwards, Vista, NetSuite, and OneStream, so the data stays open and portable rather than locked into one platform.
Frequently Asked Questions
What is Apache Iceberg?
An open table format for large analytic datasets. It sits over files in a data lake and adds database-like reliability, transactions, schema evolution, and consistent reads, which helps turn a data lake into a lakehouse.
What is the difference between Iceberg and Parquet?
Parquet is a file format that stores the data efficiently. Iceberg is a table format that organizes many Parquet files into a reliable table, adding transactions, schema evolution, and time travel. They work together: Iceberg tables are usually made of Parquet files.
Why is Apache Iceberg important for the lakehouse?
Because it gives file-based lake storage the reliability of a warehouse table, transactions, consistent reads, and schema evolution, which is what lets a lakehouse serve as a trustworthy analytics platform on open, low-cost storage.
