The article focuses on advancements in Apache Iceberg, particularly related to Change Data Capture (CDC) in versions 3 and 4. Iceberg’s core strengths lie in its immutability and snapshot capabilities, allowing for time travel within data tables. As streaming data becomes more prevalent, challenges arise, especially when dealing with constant updates that require efficient processing. One of Iceberg's main hurdles has been the handling of deletes: traditional equality deletes can slow down reads, as they force systems to repeatedly scan for changes. Version 3 introduced deletion vectors, which allow for more precise targeting of rows that need to be removed, helping to streamline reading processes. It also improved row lineage, ensuring that updates retain their identification throughout changes. Despite these improvements, the system still struggles with incremental joins across multiple tables, as it lacks a global clock to synchronize operations. This is where version 4 aims to make significant strides by simplifying metadata management and enhancing efficiency in streaming operations. Version 4 proposes a more compact manifest system, which would help eliminate some of the inefficiencies found in version 3. By decoupling manifests from partition specifications, Iceberg can offer more flexible data management. This change is expected to make small commits cheaper and improve the overall handling of change planning. The article emphasizes that while previous versions laid important groundwork for CDC, the evolving needs of streaming data call for continued refinement in how Iceberg processes updates and manages metadata.