[Eliaaazzz]

Part of #39684.

The Go SDK could only support a portable logical type by registering it in three places (schema.RegisterLogicalType, coder.RegisterSchemaProviders, and the beam.SchemaProvider glue) with hand written encoders, and none of the standard logical types from schema.proto were registered. A schema from another SDK with a beam:logical_type:date:v1 or beam:logical_type:micros_instant:v1 field failed with unknown logical type.

Registration API, in package schema:

• RegisterLogicalTypeConversion[GoT, StorageT](urn, toStorage, fromStorage) registers the URN, the Go type and the storage type, and builds the row coder for GoT from the row field coder of StorageT plus the two conversion functions.
• RegisterParameterizedLogicalTypeConversion is the same for a logical type with an argument; each argument value gets its own Go type.
• RegisterLogicalTypeCoder[GoT, StorageT](urn, enc, dec) is for logical types whose wire encoding differs from the row encoding of the storage type, such as millis_instant and decimal (following PRs).
• coder.RowFieldEncoderForType and coder.RowFieldDecoderForType expose the row field encoding of a single type, which the registration builds on. Encoding an untyped nil value or a typed nil pointer returns an error. Pointer storage types are rejected at registration because the row coder decides nullability from the Go type of the field, and interface Go types are rejected because the row coder matches every implementation of an interface and the schema matches only the interface itself.

Standard types, registered by default:

• schema.Date (beam:logical_type:date:v1), a calendar date stored as days since the epoch. Encoding requires a valid date with a year from -999999999 to 999999999.
• schema.MicrosInstant (beam:logical_type:micros_instant:v1), a time.Time stored as a row of seconds and microseconds. Encoding a value with sub microsecond precision returns an error, as in the Java SDK. Decoding requires the micros field to be in [0, 999999].

The standard types are registered with runtime.RegisterType, so pipeline graphs with these types in a transform signature serialize. The row decoder also uses a provider registered for a value type when it decodes a top level pointer to that type, as the row encoder already does.

time.Time keeps its current Go specific encoding; the standard types are separate Go types, so existing pipelines are unaffected. Because the Go row coder is built from the Go type alone, each registered logical type needs its own Go type, and a parameterized one needs a Go type per argument value.

Regression test changes in sdks/go/test/regression/coders/fromyaml:

• The expected row of a beam:coder:row:v1 case is built from the schema in the coder payload through schema.ToType, so logical type fields get their registered Go types. This replaces the name based type table.
• Coder payloads are decoded from the yaml string the same way the examples are, so payload bytes above 0x7f (used by the timestamp:v1 cases) are read correctly.
• Encoding mismatches now count as failures; before, they were only logged.
• The micros_instant case runs. The remaining logical type cases are skipped by URN with a reason pointing at [Task]: Go SDK portable logical type support #39684.

Tests: schema conversion and byte level row encoding for Date and MicrosInstant (the bytes are the micros_instant example of standard_coders.yaml), date conversion at the epoch boundaries, the sub microsecond error, RegisterLogicalTypeConversion end to end with a user type, and RowFieldEncoderForType/RowFieldDecoderForType for atomic, slice and struct types. Tests also cover interface Go type rejection, the date and micros storage ranges, a nil pointer value, a top level pointer with a value provider, and graph serialization of the standard types.

---

Thank you for your contribution! Follow this checklist to help us incorporate your contribution quickly and easily:

• Mention the appropriate issue in your description (for example: addresses #123), if applicable. This will automatically add a link to the pull request in the issue. If you would like the issue to automatically close on merging the pull request, comment fixes #<ISSUE NUMBER> instead.
• Update CHANGES.md with noteworthy changes.
• If this contribution is large, please file an Apache Individual Contributor License Agreement.

See the Contributor Guide for more tips on how to make review process smoother.

To check the build health, please visit https://github.com/apache/beam/blob/master/.test-infra/BUILD_STATUS.md

GitHub Actions Tests Status (on master branch)

See CI.md for more information about GitHub Actions CI or the workflows README to see a list of phrases to trigger workflows.

[github-actions]

Assigning reviewers:

R: @jrmccluskey for label go.

This pull request likely touches a core component ("core" label). Please review with scrutiny.

Note: If you would like to opt out of this review, comment assign to next reviewer.

Available commands:

• stop reviewer notifications - opt out of the automated review tooling
• remind me after tests pass - tag the comment author after tests pass
• waiting on author - shift the attention set back to the author (any comment or push by the author will return the attention set to the reviewers)

The PR bot will only process comments in the main thread (not review comments).

[github-actions]

Reminder, please take a look at this pr: @jrmccluskey

[codecov]

Codecov Report

❌ Patch coverage is 87.58621% with 18 lines in your changes missing coverage. Please review.
✅ Project coverage is 28.76%. Comparing base (76f6e6d) to head (e7c4f6c).
⚠️ Report is 216 commits behind head on master.

Files with missing lines	Patch %	Lines
sdks/go/pkg/beam/core/graph/coder/row_decoder.go	66.66%	4 Missing and 4 partials ⚠️
...kg/beam/core/runtime/graphx/schema/logicaltypes.go	89.83%	3 Missing and 3 partials ⚠️
sdks/go/pkg/beam/core/graph/coder/row_encoder.go	69.23%	3 Missing and 1 partial ⚠️

Additional details and impacted files

@@              Coverage Diff              @@
##             master   #40154       +/-   ##
=============================================
- Coverage     56.16%   28.76%   -27.41%     
=============================================
  Files          1124      381      -743     
  Lines        178336    71129   -107207     
  Branches       1489        0     -1489     
=============================================
- Hits         100170    20458    -79712     
+ Misses        75645    48822    -26823     
+ Partials       2521     1849      -672     

Flag	Coverage Δ
go	28.76% <87.58%> (-0.38%)	⬇️
java	?
python	?

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[jrmccluskey]

Hey there! I'm starting a review pass, if you would go ahead and rebase this PR now that the two diffbase PRs are merged that would make this a little bit smoother as far as reviewing and merging. Thanks!

[Eliaaazzz]

Thanks a lot for reviewing this series, Jack, and for merging #40152 and #40153! I rebased this one onto master, so the diff now only shows its own changes.

I also pushed a small commit that swaps reflect.PtrTo for reflect.PointerTo in row_test.go, since Yi pointed out on #40153 that PtrTo is deprecated.

Happy to adjust anything you spot.

[jrmccluskey]

A few notes from the initial pass

[jrmccluskey]

Date doesn't have a valid zero value (e.g. Date{} won't encode and will be rejected.) I would suggest treating Date{} as the epoch for clarity.

[Eliaaazzz]

Done in e7c4f6c. The zero Date now stands for the epoch, so Date{} encodes as day 0 and prints as 1970-01-01. Decoding day 0 still returns the explicit fields.

[jrmccluskey]

Now that users will be interacting with this code more directly, it's probably a good idea to include a duplication check here (I am okay with this being a follow-up.)

[Eliaaazzz]

Added in e7c4f6c. RegisterLogicalType now panics on conflicting registrations for a Go type or an identifier and argument. Registering the same logical type again has no effect.

[jrmccluskey]

Consider implementing String() and Equals() methods on this type for ease of use.

[Eliaaazzz]

Added in e7c4f6c. String prints RFC 3339 with microseconds and Equals wraps time.Time.Equal.

[Eliaaazzz]

Thanks for the notes. I pushed e7c4f6c with the three changes and merged master to clear the CHANGES.md conflict.

[jrmccluskey]

LGTM, thank you!