-
Notifications
You must be signed in to change notification settings - Fork 8
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Initial set of commits #1
Conversation
- remove timestamp_tz parameter - add uuid (exists in iceberg & trino) - add more range information for types. - move simple, compound and variations to wip status - Clarify interval ranges and change interval arrow type. Co-authored-by: Weston Pace <weston.pace@gmail.com>
* run protoc for PRs
* initial example of embedded protobuf * bump mkdocs-protobuf version
- Add boolean type and literals - Add a typed null literal - Fix map type definition - Add java option to generate multiple files.
Co-authored-by: Weston Pace <weston.pace@gmail.com>
Updates to ideally support majority of tpch queries - Remove aggregate expressions type from generalized expressions. (only allow aggregate expressions as root expressions for aggregation) - Update function mapping to support options - Remove named structs from type unions (should only be used in special places as root, not in arbitrary hierarchy) - Add project, join, fetch, aggregate, sort, set logical relational operations. - Introduce key scalar and aggregate functions in functions yaml. - Remove old extensions docs - Add nullability handling and type parsing syntax. Address substrait-io#42, substrait-io#43, substrait-io#44
…trait-io#49) * Proposed updates to support required and optional enumerations. * Apply suggestions from code review Co-authored-by: Weston Pace <weston.pace@gmail.com>
* Add List literal message type * Add list field to oneof literal
Various YAML cleanups: * Remove duplicate anchor * Make whitespace consistent * Add required space after map colon * Make spacing more consistent * Add yamllint config * Add yamllint job to PR workflow * Format aggregate_functions yaml * Format extension_types yaml * Format scalar_functions yaml * Remove unnecessary brackets * Format type_variations * Use slightly more compact repr for some types fields * Increase yaml line length restriction to 120 * Fix mkdocs.yml * Wrap description * Revert back to bracket syntax * Unwrap lines
* Overriding the C# namespace as Io.Substrait is an odd namespace (Io is not an organization) * Added Substrait.Protobuf C# namespace to more files
The join type enum is in JoinRel but not defined as an actual field in there.
* Update functions to flatter format. This closes substrait-io#65
- Change the slack invite link to be a variable within mkdocs.yml - Update the link for another 30 days. This closes substrait-io#78
* Add NamedRel * Move NamedRel out and rename to RelRoot * Fix doc * Fix numbering * Remove newline * Fix reuse bug
* Add schema for yaml extensions - update existing extensions to validate against schema - add an example "unknown" extension to illustrate simplified possibilities. * Add github job to validate yaml extensions. * fix gh workflow * Address review comments. * remove extraneous workflow op.
* Clarify/solidify extensions - Clarify extensions including documentation, updated binary representation, distinction between simple and advanced. - Provide documentation on serialization and more serialization examples, including details around surrogate keys and pointers - Update all fields that are anchors/reference to use uint32 as opposed to complex structures. Update field names to better clarify behavior - Add extension capabilities to protobuf for advanced extensions (along with supporting documentation) - Update type variations proto definition and extension definition pattern to be consistent with other extension types. - Reorder binary serialization to be first in nav given it's greater maturity over text. - Remove HintKeyValue and replace with AdvancedExtension use. Co-authored-by: Weston Pace <weston.pace@gmail.com>
* Add proto format CI check * Fix formatting Co-authored-by: Jacques Nadeau <jacques@apache.org>
* Minor expression cleanups - Add java package name for all proto files - Add an explicit CAST expression - Update the varchar literal to specify it's length - Move AggregateFunction, AggregationPhase and SortField out of the Expression message (top-level in expression.proto). - Remove AggregationPhase from aggregate rel since it is specified on the individual aggregation function bindings - Add support for a filter as part of an aggregation measure (SQL2003) - Cleanup some markdown linebreaks - Add structure for a return program (complex return operations)
…ubstrait-io#98) * Rename Switch/If expressions to reflect their use case better Also use `Literal` in case value to restrict according to the use case * Use Literal for SwitchExpression if field
…n should apply to the schema before projection (i.e. it will include fields that are removed by the projection) (substrait-io#159)
* Update mkdocs-protobuf plugin previously mkdocs-protobuf had pinned mkdocs. This patch updates the mkdocs-protobuf plugin so that mkdocs is no longer pinned and removes the workaround from substrait-io#153 * bump versions to fix build
* Remove references to organization IDs
@jvanstraten, looks like this is breaking the main build. Can you see if you can address? |
Seems unrelated to CI or anything the merge did; instead it looks like protobuf broke the world again. Google's idea of code generator vs runtime version is not dissimilar from "any color, as long as it's black," and I guess they've diverged, at least on CI. I could work around it pretty easily, but it's not going to fix the root cause, especially if we want to actually push the validator to PyPI; the generator version would be whatever is on CI, whereas the runtime version is whatever the user happens to have installed on their system, so 💥 unless the stars align. I'll have to think about this, I guess. ETA info dump:
|
Replaces substrait-io/substrait#155, refer to that thread for details.