v0.9.0-beta20

What changed

Sane exception is thrown, when attempting to open an empty database file.

Full Changelog: v0.9.0-beta19...v0.9.0-beta20

Release v0.9.0-beta19

What's Changed

Support nippy based files in a zip based database - called zippy.
This could be the new goto nippy database format, much more space efficient than "classic" ndnippy:

• 300-400% faster query speeds!
• With just 20-25% larger size than ZIP'ed EDNs
  • for comparison, ndnippy is usually as large as the uncompressed EDNs

Full Changelog: v0.9.0-beta18...v0.9.0-beta19

v0.9.0-beta18

What's Changed

• Can now use ZIPs of EDN or JSON files directly as databases. See examples in the tests.
• Cheshire has been replaced with Charred (pure clojure JSON parser)
• Updated Clojure to 1.12.0

Full Changelog: v0.9.0-beta12...v0.9.0-beta18

v0.9.0-beta12

Changes:

• ns nd-db.compress has been removed and functionality extracted into new library - com.luposlip/clarch
• new fn or-q take a seq of dbs, and queries them in order until a value is returned

Full Changelog: v0.9.0-beta10...v0.9.0-beta12

v0.9.0-beta11

Wrapping gz outputstream with tar outputstream for new fn tar-gz-outputstream.

Full Changelog: v0.9.0-beta10...v0.9.0-beta11

v0.9.0-beta10

• Add helper fns in the nd-db.compress namespace for saving (not yet reading) .tar.gz archives.
• Update dependencies

Full Changelog: v0.9.0-beta9...v0.9.0-beta10

v0.9.0-beta9

Updated dependencies, and throw sane error when nippy versions are incompatible.

Full Changelog: v0.9.0-beta8...v0.9.0-beta9

v0.9.0-beta8 - bugfix release

Fixes a bug with the index returned from appending docs.

v0.9.0-beta7 - append docs

This version makes it possible to add documents to an existing nd-db database, efficiently appending to the database file and the index (nddbmeta) file:

• Append documents to existing nd-db files (previously v1.0.0) g- Optional end-pointer parameter for versioning
  • no parameter:
    • use everything in the file, including new doc versions
    • added docs will update index (thus prevent getting new db value until done)
    • the index will contain only the newest version of each document
    • a future version of nd-db might contain historical versions
  • parameter:
    • look for nddbmeta using same line (name of index reflecting lines)
    • if nddbmeta doesn't exist, stop indexing after passed line number
    • this will create a new .nddbmeta file with a hash and metadata reflecting

Multiple documents are automatically written to db (and index) in batches of 128.

v0.9.0-beta2 - compression

Added nd-db.compress helper namespace for convenience.