flyrain commented on code in PR #6002:
URL: https://github.com/apache/iceberg/pull/6002#discussion_r1006141799
##########
core/src/main/java/org/apache/iceberg/ManifestEntry.java:
##########
@@ -118,6 +124,27 @@ default boolean isLive() {
@Deprecated
void setSequenceNumber(long sequenceNumber);
+ /**
+ * Returns the file sequence number.
+ *
+ * <p>The file sequence number represents the sequence number of the
snapshot in which the
+ * underlying file was added. The file sequence number is always assigned at
commit and cannot be
+ * provided explicitly, unlike the data sequence number. The file sequence
number does not change
+ * upon assigning and must be preserved in existing and deleted entries.
+ *
+ * <p>This method can return null if the file sequence number is unknown.
This may happen while
+ * reading a v2 manifest that did not persist the file sequence number for
manifest entries with
+ * status EXISTING or DELETED (older Iceberg versions).
+ */
+ Long fileSequenceNumber();
Review Comment:
IIUC, assuming df1 and df2 have data sequence number 1, once they are
compacted to one data file df3. df3's data sequence number will be 1(the old
one), and file sequence number is gong to be larger than 1, like 2 or 3. Is
that correct?
How do we assign the data data sequence number, if df1 and df2 has different
data sequence numbers?
```
df1 data seq no.: 1
df2 data seq no.: 2
compact df1 and df2 to df3
df3 data seq no.?
```
--
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]
