gudusoft.gsqlparser.parser

Class OceanBaseOracleSqlParser

Object

gudusoft.gsqlparser.parser.AbstractSqlParser

gudusoft.gsqlparser.parser.OceanBaseOracleSqlParser

All Implemented Interfaces:

SqlParser

public class OceanBaseOracleSqlParser
extends AbstractSqlParser

OceanBase user tenant — Oracle compatibility mode parser (Phase 3 fork).

This class is a verbatim fork of OracleSqlParser with class names rewritten to point at the forked grammar artifacts under gsp_java_parser/.../parser/oceanbaseoracle/. The structural intent is identical to the Oracle parser; what differs is:

• It uses the forked TLexerOceanbaseoracle, TParserOceanbaseoracle, and TParserOceanbaseoracleplsql (generated from the forked lzlexoceanbaseoracle.l, lzyaccoceanbaseoracle.y, and lzyaccoceanbaseoracleplsql.y sources).
• It registers as EDbVendor.dbvoceanbase so downstream components (sqlcmds, IdentifierProfile, the various TSQLEnv EnumMaps) consult OceanBase rules rather than Oracle rules.
• It is selected by OceanBaseSqlParser when the active EOBTenantMode is EOBTenantMode.ORACLE.

Phase 3 ships this fork at zero divergence from the Oracle base grammar — same shift/reduce conflicts (60 s/r + 2 r/r in SQL parser), same token positions, same statement coverage including full PL/SQL (packages, procedures, functions, triggers, anonymous blocks, SQL*Plus commands, hierarchical query, MERGE, MODEL clause). OceanBase-specific Oracle-mode extensions (hints, partition extensions, tenant DDL, outlines, tablegroups) arrive incrementally in Phase 4 via additive edits to the .y sources. Each Phase 4 addition is tracked in oceanbaseoracle/FORK_DIVERGENCE.md.

Token alignment between lzyaccoceanbaseoracle.y and lzyaccoceanbaseoracleplsql.y is mandatory and enforced by oceanbaseoracle/check_token_alignment.sh. Both files MUST declare identical tokens at identical positions or the SQL lexer and the PL/SQL parser disagree silently — see CLAUDE.md and the Phase 0 plan for the rationale.

Backporting upstream Oracle grammar fixes is documented in the Phase 0 plan (gsp_java_core/doc/oceanbase/PHASE0_REPORT.md) and the per-fork oceanbaseoracle/REGEN.md.

Since:

4.0.1.4

See Also:

the unmodified base parser this class was forked from, the mode-routing adapter that selects this parser, forked lexer, forked SQL parser, forked PL/SQL parser

Nested Class Summary

Nested classes/interfaces inherited from class gudusoft.gsqlparser.parser.AbstractSqlParser

AbstractSqlParser.PreparedSqlReader

Field Summary

Fields

Modifier and Type	Field and Description
TLexerOceanbaseoracle	flexer The Oracle lexer used for tokenization

Fields inherited from class gudusoft.gsqlparser.parser.AbstractSqlParser

defaultDelimiterStr, delimiterChar, frameStack, globalContext, globalFrame, lexer, parserContext, sourcetokenlist, sqlcmds, sqlEnv, sqlstatements, syntaxErrors, vendor

Constructor Summary

Constructors

Constructor and Description
OceanBaseOracleSqlParser() Construct Oracle SQL parser.

Method Summary

Modifier and Type	Method and Description
protected void	afterStatementParsed(TCustomSqlStatement stmt) Override to provide Oracle-specific post-processing after statement parsing.
protected void	doAfterTokenize(TSourceTokenList tokens) Post-tokenization: merge ${...} template variable tokens into single IDENT tokens.
protected void	extractVendorRawStatements(SqlParseResult.Builder builder) Call Oracle-specific raw statement extraction logic.
protected TCustomLexer	getLexer(ParserContext context) Return the Oracle lexer instance.
protected TCustomParser	getParser(ParserContext context, TSourceTokenList tokens) Return the Oracle SQL parser instance with updated token list.
protected TCustomParser	getSecondaryParser(ParserContext context, TSourceTokenList tokens) Return the Oracle PL/SQL parser instance with updated token list.
protected void	performInterpreter(ParserContext context, TStatementList statements) Perform Oracle-specific AST interpretation/evaluation using TASTEvaluator.
protected TStatementList	performParsing(ParserContext context, TCustomParser parser, TCustomParser secondaryParser, TSourceTokenList tokens, TStatementList rawStatements) Perform full parsing of statements with syntax checking.
protected void	performSemanticAnalysis(ParserContext context, TStatementList statements) Perform Oracle-specific semantic analysis using TSQLResolver.
protected void	setupVendorParsersForExtraction() Setup Oracle parsers for raw statement extraction.
protected void	tokenizeVendorSql() Call Oracle-specific tokenization logic.
String	toString() Copy syntax errors from a statement to our error list.

Methods inherited from class gudusoft.gsqlparser.parser.AbstractSqlParser

attemptErrorRecovery, copyErrorsFromStatement, doExtractRawStatements, extractRawStatements, getanewsourcetoken, getDefaultDelimiterStr, getDelimiterChar, getErrorCount, getrawsqlstatements, getSyntaxErrors, getVendor, handleStatementParsingException, initializeGlobalContext, isDollarFunctionDelimiter, onRawStatementComplete, onRawStatementCompleteVendorSpecific, parse, performTokenization, prepareSqlReader, processTokensBeforeParse, processTokensInTokenTable, setTokenHandle, tokenize, towinlinebreak

Methods inherited from class Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Field Detail

flexer

public TLexerOceanbaseoracle flexer

The Oracle lexer used for tokenization

Constructor Detail

OceanBaseOracleSqlParser

public OceanBaseOracleSqlParser()

Construct Oracle SQL parser.

Configures the parser for Oracle database with default delimiters:

• SQL statements: semicolon (;)
• PL/SQL blocks: forward slash (/)

Following the original TGSqlParser pattern, the lexer and parsers are created once in the constructor and reused for all parsing operations. This avoids unnecessary object allocation overhead since the parser is not thread-safe and designed for single-use per instance.

Method Detail

getLexer

protected TCustomLexer getLexer(ParserContext context)

Return the Oracle lexer instance.

The lexer is created once in the constructor and reused for all parsing operations. This method simply returns the existing instance, matching the original TGSqlParser pattern where the lexer is created once and reset before each use.

Specified by:

getLexer in class AbstractSqlParser

Parameters:

context - parser context (not used, lexer already created)

Returns:

the Oracle lexer instance created in constructor

getParser

protected TCustomParser getParser(ParserContext context,
                                  TSourceTokenList tokens)

Return the Oracle SQL parser instance with updated token list.

The parser is created once in the constructor and reused for all parsing operations. This method updates the token list and returns the existing instance, matching the original TGSqlParser pattern.

Specified by:

getParser in class AbstractSqlParser

Parameters:

context - parser context (not used, parser already created)

tokens - source token list to parse

Returns:

the Oracle SQL parser instance created in constructor

getSecondaryParser

protected TCustomParser getSecondaryParser(ParserContext context,
                                           TSourceTokenList tokens)

Return the Oracle PL/SQL parser instance with updated token list.

Oracle needs a secondary parser (TParserOceanbaseoracleplsql) for PL/SQL blocks (procedures, functions, packages, triggers, anonymous blocks).

The parser is created once in the constructor and reused for all parsing operations. This method updates the token list and returns the existing instance, matching the original TGSqlParser pattern.

Overrides:

getSecondaryParser in class AbstractSqlParser

Parameters:

context - parser context (not used, parser already created)

tokens - source token list to parse

Returns:

the Oracle PL/SQL parser instance created in constructor

tokenizeVendorSql

protected void tokenizeVendorSql()

Call Oracle-specific tokenization logic.

Delegates to dooraclesqltexttotokenlist which handles Oracle's specific keyword recognition, SQL*Plus commands, forward slash disambiguation, and token generation.

Specified by:

tokenizeVendorSql in class AbstractSqlParser

doAfterTokenize

protected void doAfterTokenize(TSourceTokenList tokens)

Post-tokenization: merge ${...} template variable tokens into single IDENT tokens. Template syntax like ${if(len(X) == 0, "", "...")} is used by BI tools.

Overrides:

doAfterTokenize in class AbstractSqlParser

Parameters:

tokens - the source token list (mutable)

setupVendorParsersForExtraction

protected void setupVendorParsersForExtraction()

Setup Oracle parsers for raw statement extraction.

Oracle uses dual parsers (SQL + PL/SQL), so we inject sqlcmds and update token lists for both parsers.

Specified by:

setupVendorParsersForExtraction in class AbstractSqlParser

extractVendorRawStatements

protected void extractVendorRawStatements(SqlParseResult.Builder builder)

Call Oracle-specific raw statement extraction logic.

Delegates to dooraclegetrawsqlstatements which handles Oracle's statement delimiters (semicolon and forward slash).

Specified by:

extractVendorRawStatements in class AbstractSqlParser

Parameters:

builder - the result builder to populate with raw statements

performParsing

protected TStatementList performParsing(ParserContext context,
                                        TCustomParser parser,
                                        TCustomParser secondaryParser,
                                        TSourceTokenList tokens,
                                        TStatementList rawStatements)

Perform full parsing of statements with syntax checking.

This method orchestrates the parsing of all statements by:

• Using the raw statements passed from AbstractSqlParser.parse()
• Initializing SQL and PL/SQL parsers
• Creating global context and frame stack
• Looping through each raw statement
• Calling parsestatement() on each to build AST
• Handling error recovery for CREATE TABLE/INDEX
• Collecting syntax errors

Important: This method does NOT extract raw statements - they are passed in as a parameter already extracted by AbstractSqlParser.extractRawStatements(gudusoft.gsqlparser.parser.ParserContext, gudusoft.gsqlparser.TSourceTokenList, gudusoft.gsqlparser.TCustomLexer, long). This eliminates duplicate extraction that was occurring in the old design.

Extracted from: TGSqlParser.doparse() lines 16903-17026

Specified by:

performParsing in class AbstractSqlParser

Parameters:

context - parser context

parser - main SQL parser (TParserOceanbaseoracle)

secondaryParser - PL/SQL parser (TParserOceanbaseoracleplsql)

tokens - source token list

rawStatements - raw statements already extracted (never null)

Returns:

list of fully parsed statements with AST built

afterStatementParsed

protected void afterStatementParsed(TCustomSqlStatement stmt)

Override to provide Oracle-specific post-processing after statement parsing.

For Oracle, we check if the statement is PL/SQL and recursively find syntax errors in nested PL/SQL statements.

Overrides:

afterStatementParsed in class AbstractSqlParser

Parameters:

stmt - the statement that was just parsed

performSemanticAnalysis

protected void performSemanticAnalysis(ParserContext context,
                                       TStatementList statements)

Perform Oracle-specific semantic analysis using TSQLResolver.

This includes:

• Column-to-table resolution
• Dataflow analysis
• Reference resolution
• Scope resolution

Overrides:

performSemanticAnalysis in class AbstractSqlParser

Parameters:

context - the parser context

statements - the parsed statements

performInterpreter

protected void performInterpreter(ParserContext context,
                                  TStatementList statements)

Perform Oracle-specific AST interpretation/evaluation using TASTEvaluator.

This executes simple SQL statements and evaluates expressions for static analysis and constant folding.

Overrides:

performInterpreter in class AbstractSqlParser

Parameters:

context - the parser context

statements - the parsed statements

toString

public String toString()

Copy syntax errors from a statement to our error list. Extracted from TGSqlParser.copyerrormsg().

Overrides:

toString in class AbstractSqlParser