Tags: google/zetasql
Tags
Export of internal ZetaSQL changes. -- Change by ZetaSQL Team <no-reply@google.com>: Updated the zetasql doc for the zetasql release in Sept. -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Minor refactoring against c++ resolved ast node generation templates -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by Brandon Dolphin <bdolphin@google.com>: No public description -- Change by Christoph Dibak <dibak@google.com>: Fix rewriting of public groups -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Added `ARG_TYPE_ANY_3` and `ARG_ARRAY_TYPE_ANY_3`. -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by Christoph Dibak <dibak@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by Christoph Dibak <dibak@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Refactor ResolvedAggreateScan algebrizer to remove duplicated logic -- Change by ZetaSQL Team <no-reply@google.com>: Added more tests to validate that argument aliases are not supported. -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Added `argument_alias` fields to `ResolvedFunctionArgument`. -- Change by ZetaSQL Team <no-reply@google.com>: Change edit distance implementation to use existing shared library. -- Change by ZetaSQL Team <no-reply@google.com>: Add support for null handling in the AggregateFunctionEvaluator interface -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Simplify logic for CASE expression formatting. -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Do not include internal alias in UserFacingName for table argument. -- Change by ZetaSQL Team <no-reply@google.com>: Adding utility functions `GreaterOrEqual` and `Subtract` to the rewriter utils. -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Bugfix for JSON_ARRAY_APPEND and JSON_ARRAY_INSERT. -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by Richard Levasseur <rlevasseur@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Added alias validation for function signatures. -- Change by ZetaSQL Team <no-reply@google.com>: Add test coverage for conditional evaluation in WITH queries and WITH expressions. -- Change by ZetaSQL Team <no-reply@google.com>: Add language feature to support multiway explicit UNNEST syntax -- Change by ZetaSQL Team <no-reply@google.com>: Added a new language feature FEATURE_V_1_4_ARRAY_ZIP for the array_zip function. -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Add a test case for conditional_evaluation on PIVOT and UNPIVOT -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Add test coverage for anonymization with conditional evaluation -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Replace CHECK with Status return in builtin function registry -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Added alias validation for function signatures. -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Move SetFormatValueContentOptions to SQLTestBase protected section. -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Moved `GetAliasForExpression` from the resolver code to resolved_ast_helper.h. -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Added function_signature_testutil.cc to zetasql allowlist. -- Change by ZetaSQL Team <no-reply@google.com>: Added function_signature_testutil.h into zetasql allowlist. -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Unify ReferenceDriver constructors. -- Change by ZetaSQL Team <no-reply@google.com>: Cleanup some calls to MaybeUpdateErrorMessageWithLocation() with explicit values for keep_error_location_payload. -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Add collation support for NULLIF function in the reference implementation. -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by Richard Levasseur <rlevasseur@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Add reference implementation of UDFs and templated UDFs without rewriter. -- Change by ZetaSQL Team <no-reply@google.com>: Add node_source to ResolvedArrayScan for single table array path shape -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: Clean up deprecated function signature. -- Change by ZetaSQL Team <no-reply@google.com>: Add topologically sorted generated columns information to the ResolvedInsertStmt. -- Change by ZetaSQL Team <no-reply@google.com>: Add a note about module support in docs for UDFs, TVFs, and UDAs. -- Change by ZetaSQL Team <no-reply@google.com>: No public description (And 57 more changes) GitOrigin-RevId: 2ff92e98da5ba81417caa763616253336d7d8407 Change-Id: Ib851d471d1bd3d96a36e4996ab25d5c7d381c27b
Export of internal ZetaSQL changes. -- Change by ZetaSQL Team <no-reply@google.com>: No public description -- Change by ZetaSQL Team <no-reply@google.com>: No public description GitOrigin-RevId: 818f40fee6483a896ba208ff7ee2c227a5185535 Change-Id: I9696673a553289d27c5831f9731d0641096ee934
Export of internal ZetaSQL changes. -- Change by Matthew Brown <matthewbr@google.com>: Additional RewriteVisitor tests for error paths. -- Change by Matthew Brown <matthewbr@google.com>: ZetaSQL: Fix small issues blocking release -- Change by ZetaSQL Team <no-reply@google.com>: Internal change GitOrigin-RevId: e1cd537590fec7a87d0e8b4967dcee93a711105a Change-Id: I5ebfc43120f6e0ec3f230893b5e3ac0608b68384
Export of internal ZetaSQL changes. -- Change by Matthew Brown <matthewbr@google.com>: ZetaSQL: Fix proto dependencies in ZetaSQL artifact rules -- Change by ZetaSQL Team <no-reply@google.com>: Correct a comment which implies that ResolvedAnalyticFunctionGroup.analytic_function_list only contains ResolvedAggregateFunctionCall, when it also contains ResolvedAnalyticFunctionCall. -- Change by ZetaSQL Team <no-reply@google.com>: add REPLICA as a non-reserve keyword and make it usable as a generic sub entity type. -- Change by ZetaSQL Team <no-reply@google.com>: Add minimal compliance test framework support for invoking views. -- Change by ZetaSQL Team <no-reply@google.com>: Simplify sql_test_base to eliminate multiple result sets for an individual test case. -- Change by ZetaSQL Team <no-reply@google.com>: Simplify sql_test_base to eliminate the set of feature sets implementation of test_features1 and test_features2 -- Change by ZetaSQL Team <no-reply@google.com>: Remove test_features1 compliance test option. -- Change by ZetaSQL Team <no-reply@google.com>: Remove "test_features1=" use from anonymization.test -- Change by ZetaSQL Team <no-reply@google.com>: Remove "test_features1=" use from dml_nested.test -- Change by ZetaSQL Team <no-reply@google.com>: Remove "test_features1=" use from dml_insert.test -- Change by ZetaSQL Team <no-reply@google.com>: Remove "test_features1=" use in geography_functions.test -- Change by ZetaSQL Team <no-reply@google.com>: Remove "test_features1=" use in analytic_partitionby_orderby.test -- Change by ZetaSQL Team <no-reply@google.com>: Remove "test_features1=" use in orderby_collate_queries.test -- Change by ZetaSQL Team <no-reply@google.com>: Remove "test_features1=" use in with_queries.test -- Change by ZetaSQL Team <no-reply@google.com>: Remove "test_features1=" use in serveral files where it was used with DISALLOW_GROUP_BY_FLOAT probably because `forbidden_features=` didn't exist yet. -- Change by ZetaSQL Team <no-reply@google.com>: Remove "test_features1=" use in several places related to array comparisons. -- Change by ZetaSQL Team <no-reply@google.com>: Remove "test_features1=" use in several time related test files. -- Change by ZetaSQL Team <no-reply@google.com>: Remove "test_features1=" use in several more files with only one to two uses each. -- Change by ZetaSQL Team <no-reply@google.com>: Remove "test_features1=" use for several files related to specific analytic window functions. -- Change by ZetaSQL Team <no-reply@google.com>: Flag and test infra enhancement for detailed error message for signature mismatch. -- Change by ZetaSQL Team <no-reply@google.com>: Internal Change GitOrigin-RevId: ccfcb5e9a72c160e62f407ead492830e4abeb627 Change-Id: I0b1b5317e6d534e1cce37ca14952322fd1b2d59c
Export of internal ZetaSQL changes. -- Change by Matthew Brown <matthewbr@google.com>: Support setting reserved keywords in Java LanguageOptions -- Change by Matthew Brown <matthewbr@google.com>: ZetaSQL release formatter support in java client -- Change by ZetaSQL Team <no-reply@google.com>: Never add a line break before -> in a lambda expression. -- Change by ZetaSQL Team <no-reply@google.com>: Support inferring proto type through a subquery. -- Change by ZetaSQL Team <no-reply@google.com>: Throw more explicit error when a script variable is used incorrectly -- Change by ZetaSQL Team <no-reply@google.com>: Add a test that will fail if any language feature that isn't currently applied to a test where it shouldn't be is applied to a test where it shouldn't be. This at least should slow down addition of cases that wrap language features around tests where they shouldn't be added. -- Change by ZetaSQL Team <no-reply@google.com>: Add anonymization support for TABLESAMPLE -- Change by ZetaSQL Team <no-reply@google.com>: Removing IdString::data() -- Change by ZetaSQL Team <no-reply@google.com>: Add helper functions to FunctionCallBuilder for the $case_no_value and $not functions -- Change by ZetaSQL Team <no-reply@google.com>: Used TypedList instead of TypedRange as a container for values of Range type -- Change by ZetaSQL Team <no-reply@google.com>: add helper: IntervalValue.isZero -- Change by ZetaSQL Team <no-reply@google.com>: Create a helper function to build the AggregateScan for the LIKE ANY and LIKE ALL <subquery> cases -- Change by ZetaSQL Team <no-reply@google.com>: Add ARRAY_SLICE reference implementation and compliance test -- Change by ZetaSQL Team <no-reply@google.com>: Update the ColumnFactory::MakeCol function to support collation when creating columns -- Change by ZetaSQL Team <no-reply@google.com>: Changes to the semantics of the LIKE ALL operator to comply with discussed changes when handling NULL patterns -- Change by ZetaSQL Team <no-reply@google.com>: fix bug that COLLATE() function validates number of arguments in a wrong position. -- Change by ZetaSQL Team <no-reply@google.com>: Add a KitchenSinkEnumPB type with a required enum to the test_schema -- Change by ZetaSQL Team <no-reply@google.com>: Delete a pointer to Type from TypedList since it's not used -- Change by ZetaSQL Team <no-reply@google.com>: Rollback @nullable annotation added in CL 464181150. -- Change by ZetaSQL Team <no-reply@google.com>: Add non-determinism to evaluation context for ARRAY_FILTER and ARRAY_TRANSFORM reference implementation function eval -- Change by ZetaSQL Team <no-reply@google.com>: Fix returning clause sql builder when there are table alias and column alias. -- Change by ZetaSQL Team <no-reply@google.com>: Fix bug in reference implementation of $like_any_array and $like_all_array causing errors for certain ARRAY cases -- Change by ZetaSQL Team <no-reply@google.com>: Add error messages to opaque `Preconditions.checkState()` calls in `Value`, so callers can tell what mistake they made without reading the code. -- Change by ZetaSQL Team <no-reply@google.com>: Support "RANGE" type in parser -- Change by ZetaSQL Team <no-reply@google.com>: When the parse location record type is FULL_SCOPE, make the location of an array element call include both the array expr and the index, like what the parser AST does. -- Change by ZetaSQL Team <no-reply@google.com>: Add ARRAY_SLICE reference implementation and compliance test -- Change by ZetaSQL Team <no-reply@google.com>: Documentation for WITH expression -- Change by ZetaSQL Team <no-reply@google.com>: Add ARRAY_SLICE reference implementation and compliance test -- Change by ZetaSQL Team <no-reply@google.com>: Implement range type serialization/deserialization to/from TypeProto. -- Change by ZetaSQL Team <no-reply@google.com>: Fix the default set of rewrites for the execute_query tool to be the 'ALL' set, instead of 'NONE'. -- Change by ZetaSQL Team <no-reply@google.com>: Fix regex to match string literals with ')' in parser tests -- Change by ZetaSQL Team <no-reply@google.com>: Add ARRAY_SLICE builtin function signature and rewriter -- Change by ZetaSQL Team <no-reply@google.com>: Fix NOLINT comment - START/END should be matching -- Change by ZetaSQL Team <no-reply@google.com>: Fix collation propagation on resolved column for rewriter's internal API AnalyzeSubstitute by re-designing the substitute framework. -- Change by ZetaSQL Team <no-reply@google.com>: Rename the test_outcome field of ComplianceTestCaseLables to test_error_mode so it's more clear when reading a message of this time what that field means. -- Change by ZetaSQL Team <no-reply@google.com>: fix bug that the variables inside a procedure are checked against the variables outside of the procedure. -- Change by ZetaSQL Team <no-reply@google.com>: Make ValidateTypeProto error message more clear if passed argument is empty proto -- Change by ZetaSQL Team <no-reply@google.com>: Resolve bug in FlattenRewriter that generated invalid ResolvedASTs for some query shapes. -- Change by ZetaSQL Team <no-reply@google.com>: Accommodate NULL templates in RET_CHECK() post-resolving Format() calls. -- Change by ZetaSQL Team <no-reply@google.com>: Minor change to FORMAT error handling for invalid specifiers. -- Change by ZetaSQL Team <no-reply@google.com>: remove the flag to defer folding errors to runtime. -- Change by Matthew Brown <matthewbr@google.com>: Fix parse tree for IN expression -- Change by ZetaSQL Team <no-reply@google.com>: Factor code for processing the file based test segments out of SQLTestBase. Provides a strongly typed API for SQLTestBase to interact with the [key=...] options in filebase test segments. More importantly, this consolidates the validating and parsing of those options into a single place instead of being spread all over the expansive SQLTestBase class. In some cases (such as required_features) the same option was being parsed as each time it was used. -- Change by Matthew Brown <matthewbr@google.com>: ZetaSQL: Fix RoundingMode in java client -- Change by ZetaSQL Team <no-reply@google.com>: Canonicalize CAST-ing and FORMAT-ing NaN and 0 to not be signed. -- Change by ZetaSQL Team <no-reply@google.com>: Add compliance test labels to undeclared output file -- Change by Matthew Brown <matthewbr@google.com>: ZetaSQL: Fix formatter for ZetaSQL, release basic test -- Change by Matthew Brown <matthewbr@google.com>: Fix parse tree for BETWEEN and << expressions (And 289 more changes) GitOrigin-RevId: 58a2a39ffafcb968074d8d3217b186df3a532287 Change-Id: Id3f7e303be7d73b20647de35e4756d0b9c523b74
Export of internal ZetaSQL changes.
Major changes:
We now require bazel 4.0.0
GRPC and Protobuf have been upgraded
- Older version may still work, but are not being tested
Mac builds should be improved
WORKSPACE management is different (if you require java), please
see WORKSPACE file for more info
maven_install rule is now used for java dependencies, which may
affect how bazel-based java builds work.
--
Change by Matthew Brown <matthewbr@google.com>:
Fix some spammy warnings in ZetaSQL
--
Change by ZetaSQL Team <no-reply@google.com>:
Fix interactions between joins and value tables
--
Change by ZetaSQL Team <no-reply@google.com>:
Use the KW_EXCEPT_IN_SET_OP constant rather than a string in the grammar.
--
Change by Matthew Brown <matthewbr@google.com>:
Fix macos Builds in ZetaSQL
- Clang-based builds (like the default on mac) seem to have trouble with
some of the gcc-style warning suppression - so we just spam out
all the warnings.
- Some high-frequency warnings will be cleaned up in a follow-on CL
--
Change by ZetaSQL Team <no-reply@google.com>:
Add annotation to error message when casting a STRING to INTERVAL, which is not a valid INTERVAL value.
--
Change by ZetaSQL Team <no-reply@google.com>:
Fixes SqlBuilder component for `EXCEPT DISTINCT` statements and adds a regression test.
Bug identified by GitHub user @JordyHeusdensDT and fix first proposed in #82
--
Change by ZetaSQL Team <no-reply@google.com>:
INTERNAL
--
Change by ZetaSQL Team <no-reply@google.com>:
rename tvf_bad_call_error_message to mismatch_message.
--
Change by Matthew Brown <matthewbr@google.com>:
ZetaSQL: Upgrade GRPC
--
Change by ZetaSQL Team <no-reply@google.com>:
Switch Value::Struct to Value::MakeStruct to handle errors gracefully.
--
Change by ZetaSQL Team <no-reply@google.com>:
INTERNAL
--
Change by Matthew Brown <matthewbr@google.com>:
ZetaSQL: Upgrade several java dependencies to newer versions
--
Change by ZetaSQL Team <no-reply@google.com>:
Remove some dead code.
--
Change by Matthew Brown <matthewbr@google.com>:
Stop storing a pointer to language options (make a copy instead)
--
Change by Matthew Brown <matthewbr@google.com>:
Propagate language options into parser when resolving TemplatedSqlUdf
--
Change by ZetaSQL Team <no-reply@google.com>:
Add COTH to ZetaSQL as part of adding new math functions
--
Change by ZetaSQL Team <no-reply@google.com>:
Add SECH to ZetaSQL as part of adding new math functions
--
Change by ZetaSQL Team <no-reply@google.com>:
Add CSCH to ZetaSQL as part of adding new math functions
--
Change by Matthew Brown <matthewbr@google.com>:
ZetaSQL: Require bazel 4.0.0+
--
Change by ZetaSQL Team <no-reply@google.com>:
Fix RET_CHECK failure in ResolveArrayElement related to FLATTEN.
--
Change by ZetaSQL Team <no-reply@google.com>:
Documentation for braced new constructors.
--
Change by Matthew Brown <matthewbr@google.com>:
ZetaSQL: Update protobuf version to 3.19.3
GitOrigin-RevId: 185a07d6cd6ba697ac569f3df298f9a38b24e51c
Change-Id: I454d4947229bb3aa51706923e37f55ab0919eb0b
Export of internal ZetaSQL changes. ZetaSQL now requires gcc8+ -- Change by ZetaSQL Team <no-reply@google.com>: Add collation support for COUNT(DISTINCT) in reference implementation. -- Change by ZetaSQL Team <no-reply@google.com>: Add collation support for $case_with_value function in reference implementation. -- Change by ZetaSQL Team <no-reply@google.com>: Fix zetaSQL analyzer tests that were broken by a missed include. -- Change by ZetaSQL Team <no-reply@google.com>: fix ALTER TABLE RENAME COLUMN error message while combining with other action -- Change by ZetaSQL Team <no-reply@google.com>: Refactoring function to fetch collation name from ResolvedCollation object. -- Change by ZetaSQL Team <no-reply@google.com>: Running IWYU and clean_builder on "types" project. -- Change by ZetaSQL Team <no-reply@google.com>: Rename FunctionArgumentTypeOptions::affects_function_collation -- Change by ZetaSQL Team <no-reply@google.com>: Fix internal error in SQLBuilder. -- Change by ZetaSQL Team <no-reply@google.com>: Add test case to prove HAVING clause could handle collation correctly -- Change by ZetaSQL Team <no-reply@google.com>: propagates collation through MakeResolvedCast in type casting case -- Change by ZetaSQL Team <no-reply@google.com>: Support anonymization userid column for TVFs returning value tables. -- Change by ZetaSQL Team <no-reply@google.com>: Remove deprecated and unused field `ResolvedCreateFunctionStmt::return_type` -- Change by ZetaSQL Team <no-reply@google.com>: Add a comment in all_rewriters that causes clang-format to bin-pack in a way that is less likely to cause line conflicts between changes. -- Change by ZetaSQL Team <no-reply@google.com>: Add collation support for STARTS_WITH/ENDS_WITH functions in reference implementation. -- Change by ZetaSQL Team <no-reply@google.com>: Adds a dedicated function to recursively print annotation values with a specific AnnotationSpec ID. -- Change by ZetaSQL Team <no-reply@google.com>: resolve collation for any aggregate function call with is_distinct==true -- Change by ZetaSQL Team <no-reply@google.com>: Add compliance tests for collation features with JOIN and HAVING. -- Change by ZetaSQL Team <no-reply@google.com>: change collation handling options on several built-in functions -- Change by Matthew Brown <matthewbr@google.com>: Add sql_mode to execute_query to allow evaluating sql expressions -- Change by ZetaSQL Team <no-reply@google.com>: Add test to prove that collation propagation works in qualify clause. -- Change by ZetaSQL Team <no-reply@google.com>: Adjust required features in `tablesample.test` so that compliance tests for regular `TABLESAMPLE` aren't gated on the stratified version. -- Change by ZetaSQL Team <no-reply@google.com>: Add more compliance tests for collation features with collation name being 'binary'. -- Change by ZetaSQL Team <no-reply@google.com>: Add deprecated annotation to some already deprecated functions. #cleanup -- Change by ZetaSQL Team <no-reply@google.com>: For CTAS, propagates collation from the select query output column list to column definition list. -- Change by ZetaSQL Team <no-reply@google.com>: generate ResolvedCast node when casting string collation into string without collation. -- Change by ZetaSQL Team <no-reply@google.com>: Set preserve_in_literal_remover field of second argument of COLLATE function. Also did a small cleanup for resolved_ast validator. -- Change by ZetaSQL Team <no-reply@google.com>: Convert one more `AddCastOrConvertLiteral` to `CoerceExprToType`. In this case, we are incidentally moving the error message parser location. Either location is reasonable. Before we were pointing at the `RETURNS` clause if the `CoercesTo` check failed but pointing at the function body if the literal conversion failed. Now we are pointing to the function body in both cases. -- Change by Matthew Brown <matthewbr@google.com>: Refactor replace_with_collation compliance tests; add explicit testing of 'und' -- Change by ZetaSQL Team <no-reply@google.com>: Change type and message for a few collation-related error. This would make known-error detection easier for random query tests. -- Change by ZetaSQL Team <no-reply@google.com>: Convert two more simple cases `AddCastOrConvertLiteral` to `CoerceExprToType`. -- Change by ZetaSQL Team <no-reply@google.com>: Cleanup. Resolve TODO by moving `SelectListColumn` classes from expr_resolution_helper.h to query_resolution_helper.h -- Change by ZetaSQL Team <no-reply@google.com>: Cleanup the construction of `SelectColumnState` to make it more like a class according to our syleguide. Use unique_ptr to transfer ownership and include the mostly commonly initialized fields as constructor arguments. -- Change by ZetaSQL Team <no-reply@google.com>: Add a team policy that attaches or replaces a ParseLocation to a status when propagated using RETURN_IF_ERROR or ASSIGN_OR_RETURN. There is an existing macro `RETURN_SQL_ERROR_AT_IF_ERROR` that is similar. There are two differences. The new `LocationOverride` team policy only overrides the attached parse location. It doesn't also override the error code. This makes it much less likely that its usage will cause accidental consumption of kInternal, kResourceExhausted, kUnimplemented and other error codes that the using code didn't expect to be handed. The second difference is a bit more cosmetic, we use the TeamPolicy idiom instead of defining our own macro. -- Change by ZetaSQL Team <no-reply@google.com>: Add more tests related to `SELECT AS <prototype> <literal>` -- Change by ZetaSQL Team <no-reply@google.com>: [Bug Fix] Avoid algrebrizing Empty ResolvedCollation for KeyArg AggregateOp. -- Change by ZetaSQL Team <no-reply@google.com>: Support label in GetNextStatementProperties. -- Change by ZetaSQL Team <no-reply@google.com>: Migrate calls to `AddCastOrConvertLiteral` to the more idiomatic and less likely to use incorrectly `CoerceExprToType`. This is a yak-shaving cleanup that will hopefully reduce copy-paste bugs and makes existing code more readable. -- Change by ZetaSQL Team <no-reply@google.com>: Add some more interesting corner case tests regarding `LIMIT` -- Change by ZetaSQL Team <no-reply@google.com>: Add a `GetStatementProperties` test to verify correct behavior for struct variable assignment. -- Change by ZetaSQL Team <no-reply@google.com>: Fix implementation of TO_JSON_STRING for JSON inputs when pretty_print is true. -- Change by ZetaSQL Team <no-reply@google.com>: Refactor `CoerceExprToType` to be a bit more convenient and a bit more readable. Change the boolean argument that controls coercion mode into an enum. The enum will allow more modes and makes callsites more readable. Remove the overload taking `char* clause_name` and replace it with two overloads, one taking a complete error message template (for custom error message without a lambda) and another with no error message parameter to return the generic template (the equivalent of providing a nullptr to `cluase_name` before). -- Change by ZetaSQL Team <no-reply@google.com>: Add collation support for STRPOS/INSTR functions in reference implementation. -- Change by ZetaSQL Team <no-reply@google.com>: Support several missed statement types in GetNextStatementProperties. -- Change by ZetaSQL Team <no-reply@google.com>: Update structure of signatures for aggregate functions. -- Change by ZetaSQL Team <no-reply@google.com>: Reference impl for ARRAY_INCLUDES functions. -- Change by ZetaSQL Team <no-reply@google.com>: Add a test to show behavior for window spec with literal `>=2^63` -- Change by ZetaSQL Team <no-reply@google.com>: Add no-op case in `FunctionResolver::AddCastOrConvertLiteral` and simplify callsites that previously checked for unequal types. -- Change by ZetaSQL Team <no-reply@google.com>: Bugfix. Do not override `has_explicit_type` in `ConvertLiteralToType` unless `set_has_explicit_type` is true. -- Change by ZetaSQL Team <no-reply@google.com>: Update error message when an INTERVAL literal contains an expression that is neither INT64 or string literal. -- (And 276 more changes) GitOrigin-RevId: 25c5bf17ad00c8b8f341b2cf15da3d94ec95b796 Change-Id: I8128bb928a617c1256ae766d129a311891588fa7
Export of internal ZetaSQL changes. Add support for read-modify-write of java ResolvedAST (see RewritingVisitorTest.java for examples) Add support for TO_JSON_STRING. -- Change by ZetaSQL Team <no-reply@google.com>: Pass correct language options to Validator in PreparedQuery api when validating a resolved tree passed to the PreparedQuery constructor. -- Change by ZetaSQL Team <no-reply@google.com>: Remove in_development tag from FEATURE_BIGNUMERIC_TYPE. -- Change by ZetaSQL Team <no-reply@google.com>: Increase the number of operations per benchmark loop in value_test -- Change by ZetaSQL Team <no-reply@google.com>: Update zetasql documentation to include BIGNUMERIC support. -- Change by ZetaSQL Team <no-reply@google.com>: Remove unused lexer rules. -- Change by ZetaSQL Team <no-reply@google.com>: Analyzer changes for QUALIFY clause -- Change by ZetaSQL Team <no-reply@google.com>: Fix DEFAULT column syntax error messages. -- Change by ZetaSQL Team <no-reply@google.com>: Make a clarification about the promotion of NULLs in GREATEST/LEAST -- Change by ZetaSQL Team <no-reply@google.com>: Fix bug with MODIFY_MAP rewriter - NULL first arg should be NULL output -- Change by ZetaSQL Team <no-reply@google.com>: Fix bug in handling array offsets with FLATTEN. -- Change by ZetaSQL Team <no-reply@google.com>: execute_query: Add JSON and Textproto output modes -- Change by Matthew Brown <matthewbr@google.com>: Fix name and contract for DescriptorPool.getAllFileDescriptors to be ordered -- Change by ZetaSQL Team <no-reply@google.com>: remove expected error for contains_key -- Change by ZetaSQL Team <no-reply@google.com>: Undeprecate presence tests in proto3. -- Change by ZetaSQL Team <no-reply@google.com>: INTERNAL -- Change by ZetaSQL Team <no-reply@google.com>: execute_query: Move "ExecuteQuery" overload into test code -- Change by ZetaSQL Team <no-reply@google.com>: execute_query: Retrieve examination callback only once -- Change by ZetaSQL Team <no-reply@google.com>: execute_query: Add writers for Textproto and JSON formats -- Change by Matthew Brown <matthewbr@google.com>: local_service: Use a more explicit for serializing DescriptorPools -- Change by ZetaSQL Team <no-reply@google.com>: Support parameterized types in MERGE INSERT -- Change by ZetaSQL Team <no-reply@google.com>: Add utilities to generate protobuf messages from iterator -- Change by ZetaSQL Team <no-reply@google.com>: Fix resolver logic for PIVOT to handle the case where columns in the pivot input have multiple aliases, but only some of them are referenced. -- Change by ZetaSQL Team <no-reply@google.com>: Remove deprecated AddCastOrConvertLiteral function -- Change by Matthew Brown <matthewbr@google.com>: Remove Serializable from DescriptorPool.sub interfaces -- Change by ZetaSQL Team <no-reply@google.com>: Add explicit CAST node to struct fields with type parameters -- Change by ZetaSQL Team <no-reply@google.com>: Change the visibility of cast_date_time. -- Change by ZetaSQL Team <no-reply@google.com>: Use separate control flow nodes for starting a FOR loop and advancing a FOR loop. -- Change by ZetaSQL Team <no-reply@google.com>: Fix the AST locations for TVF arguments. If it refers to an expression, just pass the location of the expression to the downstream. Otherwise it would prevent literal casting for struct fields. -- Change by ZetaSQL Team <no-reply@google.com>: Use null handling behavior enum in reference impl instead of hard-coded list of functions that respect nulls. -- Change by ZetaSQL Team <no-reply@google.com>: Modify TypeParameters:MatchType() -- Change by ZetaSQL Team <no-reply@google.com>: Automatically FLATTEN for IN UNNEST. -- Change by ZetaSQL Team <no-reply@google.com>: Native PIVOT implementation in reference impl, which doesn't have limitations on types of aggregate functions supported imposed by the rewriter. -- Change by ZetaSQL Team <no-reply@google.com>: Split long logging statements to multiple lines. -- Change by ZetaSQL Team <no-reply@google.com>: Add flag to analyze_query to control rewrites. -- Change by ZetaSQL Team <no-reply@google.com>: RQG for ARRAY_FILTER and ARRAY_TRANSFORM. -- Change by ZetaSQL Team <no-reply@google.com>: execute_query: Add option to gracefully handle statement prompt errors -- Change by ZetaSQL Team <no-reply@google.com>: Add string escaping in test_function.cc. (And 84 more changes) GitOrigin-RevId: de685b7ab6095676abb23b49ff0dfc3e51eb09d5 Change-Id: Ib21e0f71bdb8ee07c8d65d122f8663b321e60078
Export of internal ZetaSQL changes. Add support for read-modify-write of java ResolvedAST (see RewritingVisitorTest.java for examples) Add support for TO_JSON_STRING. -- Change by ZetaSQL Team <no-reply@google.com>: Pass correct language options to Validator in PreparedQuery api when validating a resolved tree passed to the PreparedQuery constructor. -- Change by ZetaSQL Team <no-reply@google.com>: Remove in_development tag from FEATURE_BIGNUMERIC_TYPE. -- Change by ZetaSQL Team <no-reply@google.com>: Increase the number of operations per benchmark loop in value_test -- Change by ZetaSQL Team <no-reply@google.com>: Update zetasql documentation to include BIGNUMERIC support. -- Change by ZetaSQL Team <no-reply@google.com>: Remove unused lexer rules. -- Change by ZetaSQL Team <no-reply@google.com>: Analyzer changes for QUALIFY clause -- Change by ZetaSQL Team <no-reply@google.com>: Fix DEFAULT column syntax error messages. -- Change by ZetaSQL Team <no-reply@google.com>: Make a clarification about the promotion of NULLs in GREATEST/LEAST -- Change by ZetaSQL Team <no-reply@google.com>: Fix bug with MODIFY_MAP rewriter - NULL first arg should be NULL output -- Change by ZetaSQL Team <no-reply@google.com>: Fix bug in handling array offsets with FLATTEN. -- Change by ZetaSQL Team <no-reply@google.com>: execute_query: Add JSON and Textproto output modes -- Change by Matthew Brown <matthewbr@google.com>: Fix name and contract for DescriptorPool.getAllFileDescriptors to be ordered -- Change by ZetaSQL Team <no-reply@google.com>: remove expected error for contains_key -- Change by ZetaSQL Team <no-reply@google.com>: Undeprecate presence tests in proto3. -- Change by ZetaSQL Team <no-reply@google.com>: INTERNAL -- Change by ZetaSQL Team <no-reply@google.com>: execute_query: Move "ExecuteQuery" overload into test code -- Change by ZetaSQL Team <no-reply@google.com>: execute_query: Retrieve examination callback only once -- Change by ZetaSQL Team <no-reply@google.com>: execute_query: Add writers for Textproto and JSON formats -- Change by Matthew Brown <matthewbr@google.com>: local_service: Use a more explicit for serializing DescriptorPools -- Change by ZetaSQL Team <no-reply@google.com>: Support parameterized types in MERGE INSERT -- Change by ZetaSQL Team <no-reply@google.com>: Add utilities to generate protobuf messages from iterator -- Change by ZetaSQL Team <no-reply@google.com>: Fix resolver logic for PIVOT to handle the case where columns in the pivot input have multiple aliases, but only some of them are referenced. -- Change by ZetaSQL Team <no-reply@google.com>: Remove deprecated AddCastOrConvertLiteral function -- Change by Matthew Brown <matthewbr@google.com>: Remove Serializable from DescriptorPool.sub interfaces -- Change by ZetaSQL Team <no-reply@google.com>: Add explicit CAST node to struct fields with type parameters -- Change by ZetaSQL Team <no-reply@google.com>: Change the visibility of cast_date_time. -- Change by ZetaSQL Team <no-reply@google.com>: Use separate control flow nodes for starting a FOR loop and advancing a FOR loop. -- Change by ZetaSQL Team <no-reply@google.com>: Fix the AST locations for TVF arguments. If it refers to an expression, just pass the location of the expression to the downstream. Otherwise it would prevent literal casting for struct fields. -- Change by ZetaSQL Team <no-reply@google.com>: Use null handling behavior enum in reference impl instead of hard-coded list of functions that respect nulls. -- Change by ZetaSQL Team <no-reply@google.com>: Modify TypeParameters:MatchType() -- Change by ZetaSQL Team <no-reply@google.com>: Automatically FLATTEN for IN UNNEST. -- Change by ZetaSQL Team <no-reply@google.com>: Native PIVOT implementation in reference impl, which doesn't have limitations on types of aggregate functions supported imposed by the rewriter. -- Change by ZetaSQL Team <no-reply@google.com>: Split long logging statements to multiple lines. -- Change by ZetaSQL Team <no-reply@google.com>: Add flag to analyze_query to control rewrites. -- Change by ZetaSQL Team <no-reply@google.com>: RQG for ARRAY_FILTER and ARRAY_TRANSFORM. -- Change by ZetaSQL Team <no-reply@google.com>: execute_query: Add option to gracefully handle statement prompt errors -- Change by ZetaSQL Team <no-reply@google.com>: Add string escaping in test_function.cc. (And 84 more changes) GitOrigin-RevId: de685b7ab6095676abb23b49ff0dfc3e51eb09d5 Change-Id: Ib21e0f71bdb8ee07c8d65d122f8663b321e60078
Export of internal ZetaSQL changes. -- Change by ZetaSQL Team <no-reply@google.com>: Adds SetToEmptyArray and SetToEmptyObject to the JSONValueRef interface. -- Change by ZetaSQL Team <no-reply@google.com>: Fix a bug in function_signature_matcher when handling proto map args. -- Change by ZetaSQL Team <no-reply@google.com>: Add test to the sql parser. -- Change by ZetaSQL Team <no-reply@google.com>: Fix PIVOT to handle optional clauses in the aggregate function calls. -- Change by Matthew Brown <matthewbr@google.com>: Fix PARSE_TIMESTAMP (and friends) %r/%l/%I/%OI on Mac -- Change by ZetaSQL Team <no-reply@google.com>: Replace compacttrie with generaltrie in parsing format elements to facilitate opensource. -- Change by ZetaSQL Team <no-reply@google.com>: Return an error for nonexistent tables for DROP CONSTRAINT when IF EXISTS is omitted. -- Change by ZetaSQL Team <no-reply@google.com>: Add a rewriter for CONTAINS_KEY. -- Change by Matthew Brown <matthewbr@google.com>: fix some invalid uses of const std::string& in parse_tree.cc -- Change by ZetaSQL Team <no-reply@google.com>: Map function rewriter bugfixes. -- Change by ZetaSQL Team <no-reply@google.com>: Mark fields as @nullable in schema forwarding TVF. -- Change by ZetaSQL Team <no-reply@google.com>: Add documentation for implicit coercion from DATE to DATETIME. -- Change by Matthew Brown <matthewbr@google.com>: add tests to AnalyzerTest for SimpleTables and sub catalogs -- Change by ZetaSQL Team <no-reply@google.com>: Resolver changes to prepare pivot scans for the rewriter: -- Change by Matthew Brown <matthewbr@google.com>: Support --mode=parse in execute_query -- Change by ZetaSQL Team <no-reply@google.com>: fixed SQL Builder to generate correct SQL when the referred struct's field has no name. -- Change by benyu <benyu@google.com>: ResolvedAST and analyzer change for CREATE VIEW v(columns...). -- Change by ZetaSQL Team <no-reply@google.com>: Add a rewriter for MODIFY_MAP. -- Change by ZetaSQL Team <no-reply@google.com>: Fix typo (fuction -> function) -- Change by ZetaSQL Team <no-reply@google.com>: Add MODIFY_MAP to analyzer. -- Change by ZetaSQL Team <no-reply@google.com>: Remove rewriter registration. Depend directly on rewriters. -- Change by ZetaSQL Team <no-reply@google.com>: Make CollectColumnRefs and CorrelateColumnRefs work with lambda. -- Change by ZetaSQL Team <no-reply@google.com>: Finish resolver support for PIVOT by supporting automatically-generated column names for literal pivot values. -- Change by ZetaSQL Team <no-reply@google.com>: Move some options-related functions to analyzer_options.h. -- Change by ZetaSQL Team <no-reply@google.com>: Check language option for QUALIFY in parser code. -- Change by ZetaSQL Team <no-reply@google.com>: Add flag for each ResolvedLiteral node, indicating whether it should be preserved by the literal remover. -- Change by ZetaSQL Team <no-reply@google.com>: Wrap compliance tests with unvalidated JSON with language feature -- Change by ZetaSQL Team <no-reply@google.com>: Implement TypeParameters java class -- Change by ZetaSQL Team <no-reply@google.com>: SQLBuilder support for PIVOT clause. -- Change by ZetaSQL Team <no-reply@google.com>: Add Name() method to the Rewriter class. -- Change by ZetaSQL Team <no-reply@google.com>: Fix bug in setting output column names of pivot expression for groupby columns. -- Change by ZetaSQL Team <no-reply@google.com>: execute_query: Use ExecuteAfterPrepare -- Change by benyu <benyu@google.com>: Add parser for CLONE DATA INTO. -- Change by ZetaSQL Team <no-reply@google.com>: Reduce debug log level. -- Change by ZetaSQL Team <no-reply@google.com>: Inner implementation of the UNPIVOT clause. -- Change by ZetaSQL Team <no-reply@google.com>: Add an option to limit max nesting levels during parsing of JSON values. -- Change by ZetaSQL Team <no-reply@google.com>: Rewriting for ARRAY_FILTER and ARRAY_TRANSFORM. -- Change by ZetaSQL Team <no-reply@google.com>: execute_query: Add utility for building interactive shells -- Change by ZetaSQL Team <no-reply@google.com>: Update error logging -- Change by ZetaSQL Team <no-reply@google.com>: Add support for Bool/Bytes/Float literals as type parameters -- Change by ZetaSQL Team <no-reply@google.com>: Add support for array index accesses along flattened expression paths. -- Change by ZetaSQL Team <no-reply@google.com>: Add Bool/Float/Bytes support to SimpleValue -- Change by ZetaSQL Team <no-reply@google.com>: Internal change (And 14 more changes) GitOrigin-RevId: bce8178316e53362c5e7b4052882c4c56ddf5ad6 Change-Id: I47e1f7c440d8795d4de09a50d0c742e2e7e6c471