anoopj commented on code in PR #17198: URL: https://github.com/apache/iceberg/pull/17198#discussion_r3582281046
########## format/spec.md: ########## @@ -700,6 +700,10 @@ The `sequence_number` field represents the data sequence number and must never c The `file_sequence_number` field represents the sequence number of the snapshot that added the file and must also remain unchanged upon assigning at commit. The file sequence number can't be used for pruning delete files as the data within the file may have an older data sequence number. The data and file sequence numbers are inherited only if the entry status is 1 (added). If the entry status is 0 (existing) or 2 (deleted), the entry must include both sequence numbers explicitly. +#### Content file uniqueness + +Within a snapshot, each live manifest entry must be uniquely defined by `file_path` across all manifest files. A snapshot with multiple live entries for the same `file_path` has undefined behavior. Writers are not required to validate uniqueness because doing so can be expensive at commit time. Review Comment: In v4, we renamed `file_path` to `location`. Since this section is version agnostic, would it make sense to phrase it in field-agnostic terms (e.g. unique by file location)? Also, in v4 this invariant holds logically but not physically: the same file can appear in multiple manifests, with all but one entry logically deleted (e.g., via MDVs). Would it be better to state this in terms of the snapshot's logical state. e.g., "a snapshot must not contain more than one live entry for the same file" rather than "across all manifest files"? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
