Add CUBRID database dialect support#575
Open
search5 wants to merge 10 commits into
Open
Conversation
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## master #575 +/- ##
==========================================
+ Coverage 92.22% 93.18% +0.96%
==========================================
Files 138 165 +27
Lines 8382 9360 +978
==========================================
+ Hits 7730 8722 +992
+ Misses 477 467 -10
+ Partials 175 171 -4 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
search5
added a commit
to search5/jet
that referenced
this pull request
Apr 2, 2026
The codecov check on PR go-jet#575 reported 53.41% patch coverage (164 lines missing), failing the 90% target. This commit addresses the gap with comprehensive unit tests and CI integration test support. Unit test additions (cubrid/ package, 78.6% -> 94.5%): - select_statement: WINDOW clause, AsTable subquery, ORDER BY with NULLS FIRST/LAST, PRECEDING/FOLLOWING window frame offsets - set_statement: INTERSECT_ALL, EXCEPT_ALL (both function and method forms), ORDER_BY/LIMIT/OFFSET chaining, AsTable on set results - insert_statement: MODELS batch insert - replace_statement: MODELS batch replace - update_statement: SET with column assignment expressions - with_statement: WITH, WITH_RECURSIVE, CTE ALIAS (new file) - lateral: LATERAL derived table with AS alias (new file) - values: VALUES table constructor with AS alias (new file) - statement: RawStatement (new file) - functions: LPAD, RPAD, CURRENT_TIME, CURRENT_TIMESTAMP, NOW - cubrid_functions: TIMEDIFF, CURRENT_USER_, DRANDOM with seed, TO_DATETIME with format, TO_TIMESTAMP with format - dialect: argumentToString for []byte, float division (/ operator) - interval_literal: INTERVALd(0), unitType.String(), intervalDebugString Unit test additions (internal/jet/, CUBRID files -> 100%): - clause_cubrid: full serialization tests for all CUBRID-specific clauses with actual data (StartWith, ConnectBy, ConnectBy NOCYCLE, OrderSiblingsBy, MergeInto, MergeUsing, MergeOn, WhenMatched UPDATE/DELETE, WhenNotMatched INSERT, ReplaceInto with/without columns, GetColumns). Also covers all lowercase serialize() wrappers. - cast: CAST expression with dialect operator override path CircleCI configuration (.circleci/config.yml): - Add cubrid/cubrid:11.4 Docker service (port 33000, DB: cubdb) - Add wait-for-CUBRID step (up to 60s with 2s interval) - Add cubrid/... to -coverpkg list in main test run - Add CUBRID integration test step running generator/cubrid/... tests with -tags integration and CUBRID_DSN env var - Merge CUBRID coverage profile into main cover.out before codecov upload
search5
force-pushed
the
cubrid-support
branch
2 times, most recently
from
8ce27df to
81a07be
Compare
- Add cubrid/ package: SQL builder dialect for CUBRID 11.x - SELECT, INSERT, UPDATE, DELETE, REPLACE INTO, MERGE INTO - Hierarchical queries (START WITH, CONNECT BY, PRIOR, LEVEL) - CUBRID-specific functions, CAST, INTERVAL, LOCK/UNLOCK - Add generator/cubrid/ package: code generator for CUBRID schema - Table/view metadata extraction, column type mapping - DSN parsing, Pool/HA connection support - Integration tests against CUBRID 11.4 - Add internal/jet/cast.go: shared CAST expression implementation - Add internal/jet/clause_cubrid.go: CUBRID-specific SQL clauses (START WITH, CONNECT BY, ORDER SIBLINGS BY, MERGE, REPLACE INTO) - Update generator/template/: add CUBRID type mappings (short, monetary, string, nchar, clob, set, multiset, etc.) - Update cmd/jet/main.go: add CUBRID as a supported source - Update go.mod: add cubrid-go driver dependency
- Remove local replace directive for cubrid-go (use published module) - Update cubrid-go to latest version with Pool/HA support - Restore go directive to 1.24.0 to match CI toolchain
v0.1.0 declared a non-existent go 1.26.1 directive which broke builds on Go 1.24.x. v0.1.1 fixes this by lowering to go 1.22.
The codecov check on PR go-jet#575 reported 53.41% patch coverage (164 lines missing), failing the 90% target. This commit addresses the gap with comprehensive unit tests and CI integration test support. Unit test additions (cubrid/ package, 78.6% -> 94.5%): - select_statement: WINDOW clause, AsTable subquery, ORDER BY with NULLS FIRST/LAST, PRECEDING/FOLLOWING window frame offsets - set_statement: INTERSECT_ALL, EXCEPT_ALL (both function and method forms), ORDER_BY/LIMIT/OFFSET chaining, AsTable on set results - insert_statement: MODELS batch insert - replace_statement: MODELS batch replace - update_statement: SET with column assignment expressions - with_statement: WITH, WITH_RECURSIVE, CTE ALIAS (new file) - lateral: LATERAL derived table with AS alias (new file) - values: VALUES table constructor with AS alias (new file) - statement: RawStatement (new file) - functions: LPAD, RPAD, CURRENT_TIME, CURRENT_TIMESTAMP, NOW - cubrid_functions: TIMEDIFF, CURRENT_USER_, DRANDOM with seed, TO_DATETIME with format, TO_TIMESTAMP with format - dialect: argumentToString for []byte, float division (/ operator) - interval_literal: INTERVALd(0), unitType.String(), intervalDebugString Unit test additions (internal/jet/, CUBRID files -> 100%): - clause_cubrid: full serialization tests for all CUBRID-specific clauses with actual data (StartWith, ConnectBy, ConnectBy NOCYCLE, OrderSiblingsBy, MergeInto, MergeUsing, MergeOn, WhenMatched UPDATE/DELETE, WhenNotMatched INSERT, ReplaceInto with/without columns, GetColumns). Also covers all lowercase serialize() wrappers. - cast: CAST expression with dialect operator override path CircleCI configuration (.circleci/config.yml): - Add cubrid/cubrid:11.4 Docker service (port 33000, DB: cubdb) - Add wait-for-CUBRID step (up to 60s with 2s interval) - Add cubrid/... to -coverpkg list in main test run - Add CUBRID integration test step running generator/cubrid/... tests with -tags integration and CUBRID_DSN env var - Merge CUBRID coverage profile into main cover.out before codecov upload
- Remove cubridREGEXPLIKEoperator/cubridNOTREGEXPLIKEoperator from dialect.go: these operator overrides were never invoked because REGEXP_LIKE uses the dialect's RegexpLike() method path, not the binary operator override path. The default fallback in dialectImpl.RegexpLike() already outputs correct CUBRID SQL. - Add UPDATE and SHARE RowLock constants to select_statement.go - Add TestSelectFOR covering SELECT ... FOR UPDATE/SHARE - Add TestSelectPRECEDING_FOLLOWING_UNBOUNDED covering toJetFrameOffset UNBOUNDED branch cubrid package unit test coverage: 94.3% → 97.7%
Add tests covering previously uncovered code paths: - dialect.go: ValuesDefaultColumnName, JsonValueEncode lambdas - dialect.go: TRUNC function - insert_statement.go: ON DUPLICATE KEY UPDATE with multiple assignments - select_statement.go: WINDOW().AS() with no arguments, VALUES with default column names - interval_literal.go: INTERVALd hours+microseconds branch - generator/template/process.go: insertedRowAlias CUBRID branch cubrid package unit test coverage: 97.7% → 99.3%
- column_types_test.go: add DEC, CHARACTER, UTIME, VARBIT, LIST, DATETIMELTZ, TIMESTAMPLTZ, CHARACTER VARYING, and default-branch cases - error_path_test.go: unit tests for GenerateDSN (missing DB name), GeneratePool (empty DSN), GenerateHA (invalid DSN), Generate (connection refused) — covers all generator function error paths without requiring a DB - query_set_test.go: mock SQL driver tests for GetTablesMetaData (list error, view error, column error via partial mock), GetEnumsMetaData (nil return), getColumnMetaData (list columns error) Unit-test-only coverage: generator/cubrid 31.9% → 66.4% Remaining success-path lines are covered by the existing integration tests
error_path_test.go: - TestGenerateDSN_ConnectionRefused: covers openConnection error in GenerateDSN - TestGenerateDB_SchemaError: covers metadata.GetSchema error path in GenerateDB - TestGenerateDB_ProcessSchemaError: covers template.ProcessSchema error path using schemaDriver (returns one mock table) + /dev/null as destDir to force fs error - Added schemaDriver and emptyDriver mock SQL drivers query_set_test.go: - TestGetColumnMetaData_Success: covers pkSet + columns construction loops using colsPksDriver (returns one column + one PK) - TestGetColumnMetaData_PKsError: covers ListPrimaryKeys error path using pksErrDriver Coverage summary (unit tests only): - generator/cubrid/query_set.go: 0% → 100% - generator/cubrid/cubrid_generator.go: 31.9% → 89.4% (remaining ~10% are success paths covered by integration tests)
Refactor cubrid_generator.go to inject dependencies via package-level variables:
- sqlOpen: replaces sql.Open("cubrid", ...) in openConnection
- newPool: replaces cubriddriver.NewPool
- newHACluster: replaces cubriddriver.NewHACluster
- poolProvider/haProvider interfaces allow mock implementations in tests
New tests in error_path_test.go:
- TestOpenConnection_Success: success path via sqlOpen mock
- TestOpenConnection_SqlOpenError: sqlOpen failure path
- TestGenerate_Success / TestGenerateDSN_Success: full happy paths
- TestGeneratePool_Success / TestGenerateHA_Success: via mockPool/mockHA
- TestGenerateDB_CustomTemplate: covers len(templates)>0 branch
generator/cubrid unit coverage: 89.4% → 100%
Tests call the SerializerFunc returned by each operator with 0 arguments, triggering the len(expressions) < 2 panic guard. cubrid/dialect.go unit coverage: 93.18% → 100%
Author
|
Hi @go-jet! 👋 Friendly ping — this PR adds first-class CUBRID dialect support. CUBRID is widely used in Korean public-sector and enterprise environments, and this would make go-jet the first type-safe SQL builder for Go to support it natively. All tests pass (including integration tests against CUBRID 11.4), coverage improved from 92.22% to 93.18%, and existing dialects are completely untouched. Would love to get your feedback whenever you have a moment. Thank you! |
Owner
|
Thanks, @search5, for the contribution. CUBRID looks like an interesting database, and I That said, I do not currently plan to add support for new dialects in Jet. Due CUBRID also still seems to be a fairly niche database, so adding and maintaining For these reasons, I do not think this is something I can review/merge at the moment, |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Motivation
CUBRID is an open-source relational database widely adopted in Korean public-sector and enterprise environments. Despite its maturity (the current stable release is 11.4), there is no type-safe SQL builder for Go that supports CUBRID natively. Developers working with CUBRID in Go are currently limited to raw SQL strings, which sacrifices the compile-time safety and composability that go-jet provides for PostgreSQL, MySQL, and SQLite.
This PR adds first-class CUBRID support to go-jet, following the same architecture and conventions used by the existing dialects.
What This PR Contains
1.
cubrid/— SQL Builder Dialect PackageA complete SQL builder dialect for CUBRID, including:
SELECT,INSERT,UPDATE,DELETEREPLACE INTO,MERGE INTOSTART WITH,CONNECT BY/CONNECT BY NOCYCLE,ORDER SIBLINGS BYPRIOR(),LEVEL,ROWNUM,CONNECT_BY_ROOT(),CONNECT_BY_ISLEAF,SYS_CONNECT_BY_PATH()NVL,NVL2,DECODE,TO_CHAR,TO_DATE,TO_DATETIME,TRUNC,INCR,DECR, etc.CASTwith CUBRID-native type targetsINTERVALliteral expressionsLOCK/UNLOCKtable statementsSET(UNION, INTERSECT, EXCEPT) operationsWITH(CTE) support2.
generator/cubrid/— Code GeneratorDBConnectionstruct, DSN string, and Pool/HA configurations viacubrid-godriver3.
internal/jet/— Minimal Core Additionscast.go: SharedCastimplementation reusable across dialectsclause_cubrid.go: CUBRID-specific SQL clauses (ClauseStartWith,ClauseConnectBy,ClauseOrderSiblingsBy,ClauseMergeInto,ClauseMergeUsing,ClauseMergeOn,ClauseWhenMatched,ClauseWhenNotMatched,ClauseReplaceInto)4.
generator/template/— Type Mapping Extensionsmodel_template.goandsql_builder_template.go:short,monetary,string,nchar,nchar varying,clob,set,multiset,sequence,list,object,oid,timestampltz,datetimeltz,datetimetz"CUBRID"case toinsertedRowAliasinprocess.go5.
cmd/jet/main.go— CLI Entry Pointcase "cubrid"to the jet CLI generator, supporting both DSN and host/port connection parameters6.
go.modgithub.com/search5/cubrid-godriver dependencyDesign Decisions
mysql/,postgres/,sqlite/,qrm/, and existinginternal/jet/files are completely untouched. All CUBRID support is purely additive.CONNECT BY) andMERGE INTOare not afterthoughts — they have dedicated clause types, proper serialization, and full test coverage.Test Plan
go build ./...passesgo vet ./...passes (excluding pre-existingWriteBytesignature warning)go test ./cubrid/— all SQL builder unit tests passgo test ./generator/cubrid/— all unit tests passgo test -tags integration ./generator/cubrid/— all integration tests pass against CUBRID 11.4go test ./internal/jet/— existing core tests still passgo test ./generator/template/— template tests pass including new CUBRID type cases