You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
FITS Binary tables store variable-length data on a heap, recording only heap pointers (location and size) in the main table. Because of that, heaps can get bloated if data is changed (e.g. by leaving some or all of old data on the heap that is no longer referenced). And, depending on the order that tables were build (e.g. column-by-column), the heap might not be optimized for reading e.g. row-by-row.
Thus, it is desirable to add a way to rebuild (defragment) the heap before writing binary tables to an output to avoid bloating and to optimized heap structure.
Support for heap defragmenting is added in PR #442
The text was updated successfully, but these errors were encountered:
FITS Binary tables store variable-length data on a heap, recording only heap pointers (location and size) in the main table. Because of that, heaps can get bloated if data is changed (e.g. by leaving some or all of old data on the heap that is no longer referenced). And, depending on the order that tables were build (e.g. column-by-column), the heap might not be optimized for reading e.g. row-by-row.
Thus, it is desirable to add a way to rebuild (defragment) the heap before writing binary tables to an output to avoid bloating and to optimized heap structure.
Support for heap defragmenting is added in PR #442
The text was updated successfully, but these errors were encountered: