metamorphic: ignore rangedels for deduplicating point prefixes #4280
Conversation
|
This change is |
There was a problem hiding this comment.
Reviewable status: 0 of 3 files reviewed, 2 unresolved discussions (waiting on @jbowens and @sumeerbhola)
metamorphic/key_manager.go line 31 at r1 (raw file):
// types. func (m *keyMeta) all(typs ...OpType) bool { return slices.ContainsFunc(m.history, func(i keyHistoryItem) bool {
shouldn't there be a ! here? We're returning true if one of the items in history is not one of typs
metamorphic/testdata/key_manager line 505 at r1 (raw file):
[external1 = batch1.NewExternalObj()] [db1.IngestExternalFiles(external1, "a" /* start */, "z" /* end */, "" /* syntheticSuffix */, "" /* syntheticPrefix */)] can singledel on db1: "foo@5"
[nit] add trailing newline
jbowens
force-pushed
the
metamorphic-rangedel-obj
branch
from
7acd4aa to
c1cce06
Compare
There was a problem hiding this comment.
Reviewable status: 0 of 3 files reviewed, 2 unresolved discussions (waiting on @RaduBerinde and @sumeerbhola)
metamorphic/key_manager.go line 31 at r1 (raw file):
Previously, RaduBerinde wrote…
shouldn't there be a
!here? We're returning true if one of the items inhistoryis not one oftyps
lol thanks, I really confused myself with the repeated negation. I added some local vars that maybe help.
There was a problem hiding this comment.
Reviewable status: 0 of 3 files reviewed, 3 unresolved discussions (waiting on @jbowens and @sumeerbhola)
metamorphic/key_manager.go line 31 at r2 (raw file):
// types. func (m *keyMeta) all(typs ...OpType) bool { anyNotOfProvidedType := slices.ContainsFunc(m.history, func(i keyHistoryItem) bool {
[int] containsOtherTypes to avoid the double negation.
When constructing an external object, only the first point key with each unique prefix is added to the sstable. The key manager strives to maintain consistency with this behavior to ensure that the it maintains correct key histories after an ingestion of an external object. The key manager however also will discretize range deletions, sometimes creating keyMetas for keys that do not exist as point keys on the associated writer. The presence of these was confusing the key manager's calculation of which keys would end up in the external object. This commit updates the external object calculation to account for the existence of the range deletions. Future work may refactor the tracking of range deletions to track range deletions separately from point keys and as continuous spans. Fix cockroachdb#4267.
jbowens
force-pushed
the
metamorphic-rangedel-obj
branch
from
c1cce06 to
0e84b9a
Compare
When constructing an external object, only the first point key with each unique prefix is added to the sstable. The key manager strives to maintain consistency with this behavior to ensure that the it maintains correct key histories after an ingestion of an external object. The key manager however also will discretize range deletions, sometimes creating keyMetas for keys that do not exist as point keys on the associated writer. The presence of these was confusing the key manager's calculation of which keys would end up in the external object.
This commit updates the external object calculation to account for the existence of the range deletions. Future work may refactor the tracking of range deletions to track range deletions separately from point keys and as continuous spans.
Fix #4267.