opensource

the open source the data and programming communities run on, and what it got: a 5,000-times speed-up in apache arrow, a 1,575-times throughput gap in the databricks go sdk made public, streaming for spark’s go client, the first databricks driver for golang-migrate, a change merged by the creator of elixir. by name:

  • apache spark · streaming rows and arrow records in the go client; the bugs found on the way
  • apache arrow · a 5,000-times performance bug in the go library, found, documented, fixed
  • databricks · 1,575 times the throughput of the go sdk; a merged fix to the terraform provider’s cross-account policy
  • golang-migrate · the first databricks sql warehouse driver
  • gocraft/work · the go background-jobs library, kept building
  • go (language) · a go 2 proposal, and the libraries
  • elixir (language) · merged by josé valim; ecto and phoenix
  • delta lake · a merge that evolved the schema when it should not have
  • duckdb · nested json rows out of order on the way to parquet
  • parquet-go · a panic and a corruption in the go parquet reader
  • lake-orm · an orm for the lakehouse in go, over delta and iceberg
  • nordlayer-nix · nordsecurity’s business vpn, running on nixos