Notes on Sorted Data
8 days ago
- #data-structures
- #serialization
- #sorting
- Challenges in storing and comparing sorted data using byte strings.
- Fixed-width integers can waste space and have endianness issues.
- Variable-length integers (varints) can break ordering when compared byte-by-byte.
- Length-prefixing integers preserves ordering but uses an extra byte.
- Optimizing length-prefix with 4 bits to save space, with limitations.
- Handling signed integers by remapping to unsigned for correct ordering.
- Floating point numbers require special handling due to their binary representation.
- Length-prefixing strings can lead to incorrect ordering comparisons.
- Using null terminators for strings and composite data to preserve ordering.
- Composite data (tuples, structs) can be separated by null terminators for correct ordering.