Work with huge numbers of small files more quickly.
Safely split a large tar archive into a specified number of smaller tar archives
i - input tar archive that you want to split
o - output patternp - number of smaller archives to split the input archive intotarsplitter -m split -i archive.tar -o /tmp/archive-parts -p 4
archive.tar is 529479680 bytes, splitting into 4 parts of 132369920 bytes
First new archive is /tmp/archive-parts0.tar
Processed files= 10000
Processed files= 20000
Processed files= 30000
Initialized next tar archive /tmp/archive-parts1.tar
Processed files= 40000
Processed files= 50000
Processed files= 60000
Initialized next tar archive /tmp/archive-parts2.tar
Processed files= 70000
Processed files= 80000
Processed files= 90000
Initialized next tar archive /tmp/archive-parts3.tar
Processed files= 100000
Processed files= 110000
Processed files= 120000
Done reading input archive
All done
Multitheaded Archiving
i - input file matching pattern
o - output tar file pathp - number of threads or files to read at once. -p 10 will read up to 10 files from disk at a timetarsplitter -m archive -i folder/*.json -o archive.tar
It is possible to split large files, such as tar archives, into parts using the split utility. But you need to do a little work to precompute the byte split if you want a specific number of sub-files.
split -b 100m archive.tar
Worse, split won't keep all the files intact. Files will be split on the line, right near byte split, span archives, possibly making the archive unusable.
tarsplitter will not leave any broken files between the split archives.
It can be very slow to archive millions of small files. The tarsplitter -m archive mode will use as many cores as you want to create a tar archive, rather than the single threaded regular tar command.
See the LICENSE file in this repository.
$ claude mcp add tarsplitter \
-- python -m otcore.mcp_server <graph>