Apache Iceberg Under the Hood — Module 03: Reads, Deletes & CoW vs MoR
datainternals
0:00 / 0:00
Apache Iceberg Under the Hood — Module 03: Reads, Deletes & CoW vs MoR
9 просмотров · 2 нед. назад
datainternals
33 подписчика
9 просмотров · 2 нед. назад
Delete the same single row from two identical Iceberg tables and one rewrites a whole data file while the other writes a tiny tombstone beside it. This module shows both on screen, explains where position and equality deletes physically go, and corrects the mistake almost everyone makes — merge-on-read is not the fast option, it is the fast-WRITE option, and someone pays later.
Chapters:
00:00 One row, two very different deletes
00:53 Deleting from a file you cannot modify
02:28 The same delete, merge-on-read
03:06 Position deletes, equality deletes, and the bill
05:23 Exercise
06:10 Solution
08:16 Recap
Run the lab yourself:
• Repo: https://github.com/venkatdwe/iceberg-...
• This module's lab: https://github.com/venkatdwe/iceberg-...
• Two commands: just lab-up && just m03-deletes
Full course playlist: • Apache Iceberg Under the Hood
GitHub repo: https://github.com/venkatdwe/iceberg-...
#ApacheIceberg #DataEngineering #Lakehouse #DataLake #TableFormat