Found some old shapefiles lying around that I didn’t want to delete, but was sure I wouldn’t need in the foreseeable future, so I decided to compress them for archiving. But what format is most efficient? Here’s my little unofficial test. Any methods that took longer than I felt like waiting were disqualified (i.e. 7zip Ultra). All commands were run via the command line and timed using the time command.
While tempting to use as a generalizable compression test, the poor showing by bzip2 (one of the newest and most efficient formats) shows that it really does depend what you’re compressing.
Original file: US_tract_1980.shp Original size: 100.2 MB
| Format | Time | Size |
|---|---|---|
| 7zip Ultra | 0m56.212s | 46.4 MB |
| 7zip | 0m43.357s | 46.9 MB |
| Stuffit SitX LZ-arithmetic | 2m6.793s | 62.1 MB |
| Stuffit SitX LZ-Huffman -l 16 | 0m29.791s | 77.7 MB |
| Stuffit Gzip -l 16 | 0m31.854s | 77.7 MB |
| Gzip –fast | 0m9.980s | 78.3 MB |
| Zip | 0m10.923s | 78.5 MB |
| Gzip | 0m13.148s | 78.5 MB |
| Gzip –best | 0m13.612s | 78.5 MB |
| Stuffit Gzip | 0m22.865s | 78.8 MB |
| Stuffit SitX LZ-Huffman | 0m19.847s | 78.5 MB |
| Stuffit SitX -l 16 | 1m17.917s | 79.6 MB |
| Stuffit SitX | 1m19.718s | 79.6 MB |
| Stuffit SitX -l 1 | 1m20.588s | 79.6 MB |
| Stuffit SitX Blocksort | 0m57.750s | 83.4 MB |
| Stuffit SitX PPM | 2m25.372s | 83.4 MB |
| Bzip2 | 0m42.309s | 87.2 MB |
| Stuffit Bzip2 | 0m32.637s | 87.9 MB |
Table is sorted by size, then by time.
*7zip ultra: 7z a -t7z -m0=lzma -mx=9 -mfb=64 -md=32m -ms=on
Wow! 7zip wins by a mile! Nothing else even comes close…
Powered by FireStats
Recent Comments