RequestLimit & more fixes

This release includes a small new feature and an important bug fix.

• New feature: RequestLimit for Query and Scan for finer control of internal pagination (#231)
• Bug fix: Panic on saving struct with pointer to anonymous struct (#233)

Better recursive type support

• Support nested recursion in types (A → B → A and so on) (#227)

Encoding fixes

This release fixes these issues:

• Encoding recursive structs panics since 1.22.0 #227
• Encoding Issue: unsupported type structpb.isValue_Kind #229
• Fix BatchGet.From and FromRange with different table

Many thanks to everyone who reported the issues.
Note: v1.22.0 has been retracted. Please upgrade to the latest version if you're on it.

Fix decoding pointer fields to slices

This is a bugfix release for the following issues related to the new encoding in v1.22.0:

• Scan All Repeats First Element #228
• Encoding slices is broken on 1.22 when using structs pointer fields #230

Please note the following issues are still around:

• Recursive structs breaking #227 (fixed in 1.22.2)
• Encoding Issue: unsupported type structpb.isValue_Kind #229 (fixed in 1.22.2)

Faster encoding and multi-table batches

What's Changed

• Optimize encoding by @guregu in #224
  • Much faster encoding and decoding, see above for benchmarks (~1.5x faster, ~2x less allocs)
• Support multiple tables in BatchWrite and BatchGet by @guregu in #226
  • (*BatchWrite).PutIn for putting items in other tables
  • (*BatchWrite).DeleteIn and (*BatchWrite).DeleteInRange for deleting items in other tables
  • (*BatchWrite).Merge for merging batch writes together
  • (*BatchGet).From and (*BatchGet).FromRange for getting items from other tables
  • (*BatchGet).Merge for merging batch gets together
  • (*BatchGet).ProjectTable for specifying table-specific projections
  • (*BatchGet).IterWithTable for creating an iterator that keeps track of which table each item comes from
• Add automated testing against DynamoDB local using GitHub Actions
  • Integration tests are easier to run now, see README

Full Changelog: v1.21.0...v1.22.0

Bonus: v2 preview available

dynamo v2 will target aws-sdk-go v2 and is (finally) nearing completion.
You can test it by depending on the v2-dev branch:

go get github.com/guregu/dynamo/v2@v2-dev

Feedback is welcome at #206.

Encoding optimization (pre)

Includes #224 which greatly improves the speed and memory usage of marshaling/unmarshaling. See benchmarks in #224, or try running them locally.
If you've got a test suite that uses dynamo, it would help me a lot if you run this pre-release against it and let me know if you find anything weird.
Thanks and please look forward to a release soon. I've been running this on one of my services for a little while and it's looking good so far.

Configurable retrying

What's Changed

• Configuration for retrying by @guregu in #223 (issue: #222)
  • Depending on your aws.Config setup, the behavior for retrying can change. MaxRetries and Retryer fields will be respected now.
  • See docs: https://github.com/guregu/dynamo#retrying

Full Changelog: v1.20.2...v1.21.0

v1.20.2

Bugfix release.

• Fix nil pointer panic in DescribeTable for LSIs in Global Tables (#220)
  • Thanks @chyi-code for the bug report

Full Changelog: v1.20.1...v1.20.2

ErrTooMany + Limit(1) fix

This is a minor bugfix release.

• Don't return ErrTooMany in Query.One when Limit is set to 1
• Update dependencies

Full Changelog: v1.20.0...v1.20.1

Parallel Scans

This release adds new methods for Scan to create parallel iterators (#20): Scan.IterParallel, Scan.AllParallel and paginating versions. The parallel iterators handle the concurrency and segmenting of parallel scan requests automatically.

What's Changed

• Add parallel scan by @guregu in #219
• Use context.Context instead of aws.Context (which is a type alias to context.Context)
• Update dependencies

Full Changelog: v1.19.0...v1.20.0