Skip to content

Dgraph v20.03.0

Compare
Choose a tag to compare
@danielmai danielmai released this 31 Mar 00:42
· 1598 commits to master since this release

Note: This release requires you to export and re-import data prior to upgrading or rolling back. The underlying data format has been changed.

Changed

  • Report GraphQL stats from alpha. (#4607)
  • During backup, collapse split posting lists into a single list. (#4682)
  • Optimize computing reverse reindexing. (#4755)
  • Add partition key based iterator to the bulk loader. (#4841)
  • Invert s2 loop instead of rebuilding. (#4782)
  • Update Badger Version. (#4935)
  • Incremental Rollup and Tablet Size Calculation. (#4972)
  • Track internal operations and cancel when needed. (#4916)
  • Set version when rollup is called with no splits. (#4945)
  • Use a different stream writer id for split keys. (#4875)
  • Split posting lists recursively. (#4867)
  • Enterprise features
    • Breaking changes
      • [BREAKING] Underlying schema for ACL has changed. Use the upgrade tool to migrate to the new data format. (#4725)

Added

  • Add GraphQL API for Dgraph accessible via the /graphql and /admin HTTP endpoints on Dgraph Alpha. (#933)
  • Add support for sorting on multiple facets. Fixes #3638. (#4579)
  • Expose Badger Compression Level option in Bulk Loader. (#4669)
  • GraphQL Admin API: Support Backup operation. (#4706)
  • GraphQL Admin API: Support export, draining, shutdown and setting lrumb operations. (#4739)
  • GraphQL Admin API: duplicate /health in GraphQL /admin (#4768)
  • GraphQL Admin API: Add /admin/schema endpoint (#4777)
  • Perform indexing in background. (#4819)
  • Basic Sentry Integration - Capture manual panics with Sentry exception and runtime panics with a wrapper on panic. (#4756)
  • Ludicrous Mode. (#4872)
  • Enterprise features
    • ACL: Allow users to query data for their groups, username, and permissions. (#4774)
    • ACL: Support ACL operations using the admin GraphQL API. (#4760)
    • ACL: Add tool to upgrade ACLs. (#5016)

Fixed

  • Avoid running GC frequently. Only run for every 2GB of increase. Small optimizations in Bulk.reduce.
  • Check response status when posting telemetry data. (#4726)
  • Add support for $ in quoted string. Fixes #4695. (#4702)
  • Do not include empty nodes in the export output. Fixes #3610. (#4773)
  • Fix Nquad value conversion in live loader. Fixes #4468. (#4793)
  • Use /tmp dir to store temporary index. Fixes #4600. (#4766)
  • Properly initialize posting package in debug tool. (#4893)
  • Fix bug, aggregate value var works with blank node in upsert. Fixes #4712. (#4767)
  • Fix count with facets filter. Fixes #4659. (#4751)
  • Change split keys to have a different prefix. Fixes #4905. (#4908)
  • Various optimizations for facets filter queries. (#4923)
  • Throw errors returned by retrieveValuesAndFacets. Fixes #4958. (#4970)
  • Add "runInBackground" option to Alter to run indexing in background. When set to true, then the Alter call returns immediately. When set to false, the call blocks until indexing is complete. This is set to false by default. (#4981)
  • Set correct posting list type while creating it in the live loader. Fixes #4889. (#5012)
  • Breaking changes
    • [BREAKING] Language sorting on Indexed data. Fixes #4005. (#4316)