Reverse scans are currently much slower than forward scans:
Dataset Size: 50,000 items
KeyFormat: String (up to 10 chars)
ValueFormat: String (up to 10 chars)
Read only scenario
Insertion of 50k items takes: 372 msec
Forward Full table scan (just call cursor->first and then cursor->next until end): 9 msec
Backward Full table scan (just call cursor->first and then cursor->next until end): 46,600 msec
- related to
-
WT-82 process_value() doesn't handle bad values
- Closed
-
WT-202 Snapshot locking
- Closed
-
WT-206 examples/python/ex_access.py is broken
- Closed
-
WT-338 Format test failure
- Closed
-
WT-351 Checkpoint the metadata after successful schema-level operations.
- Closed
-
WT-431 Read Reverse degradation
- Closed
-
WT-463 Make run-time statistics optional.
- Closed
-
WT-467 reserved identifier violation
- Closed
-
WT-469 Changes to minimize the impact of checkpoints on other threads
- Closed