oushu/lava-oushudb-dt-sql-parser
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

feat: improve hive keywords rule (#154)

Browse Source
This commit is contained in:
Hayden
2023-09-05 14:27:21 +08:00
committed by GitHub
parent 79b66ced4f
commit ded9f28e36
10 changed files with 12493 additions and 11060 deletions
Download Patch File Download Diff File Expand all files Collapse all files

39
src/grammar/hive/HiveSqlLexer.g4
View File

@ -92,13 +92,16 @@ KW_DATACONNECTOR : 'CONNECTOR';
KW_DATACONNECTORS : 'CONNECTORS';
KW_DATE : 'DATE';
KW_DATETIME : 'DATETIME';
KW_DAY : 'DAY' 'S'?;
KW_DAY : 'DAY';
KW_DAYS : 'DAYS';
KW_DAYOFWEEK : 'DAYOFWEEK';
KW_DBPROPERTIES : 'DBPROPERTIES';
KW_DCPROPERTIES : 'DCPROPERTIES';
KW_DDL : 'DDL';
KW_DEBUG : 'DEBUG';
KW_DECIMAL : 'DEC' 'IMAL'? | 'NUMERIC';
KW_DEC : 'DEC';
KW_DECIMAL : 'DECIMAL';
KW_NUMERIC : 'NUMERIC';
KW_DEFAULT : 'DEFAULT';
KW_DEFERRED : 'DEFERRED';
KW_DEFINED : 'DEFINED';
@ -110,7 +113,8 @@ KW_DESCRIBE : 'DESCRIBE';
KW_DETAIL : 'DETAIL';
KW_DIRECTORIES : 'DIRECTORIES';
KW_DIRECTORY : 'DIRECTORY';
KW_DISABLE : 'DISABLE' 'D'?;
KW_DISABLE : 'DISABLE';
KW_DISABLED : 'DISABLED';
KW_DISTINCT : 'DISTINCT';
KW_DISTRIBUTE : 'DISTRIBUTE';
KW_DISTRIBUTED : 'DISTRIBUTED';
@ -120,7 +124,8 @@ KW_DROP : 'DROP';
KW_DUMP : 'DUMP';
KW_ELEM_TYPE : '$ELEM$';
KW_ELSE : 'ELSE';
KW_ENABLE : 'ENABLE' 'D'?;
KW_ENABLE : 'ENABLE';
KW_ENABLED : 'ENABLED';
KW_END : 'END';
KW_ENFORCED : 'ENFORCED';
KW_ESCAPED : 'ESCAPED';
@ -161,7 +166,8 @@ KW_GROUP : 'GROUP';
KW_GROUPING : 'GROUPING';
KW_HAVING : 'HAVING';
KW_HOLD_DDLTIME : 'HOLD_DDLTIME';
KW_HOUR : 'HOUR' 'S'?;
KW_HOUR : 'HOUR';
KW_HOURS : 'HOURS';
KW_IDXPROPERTIES : 'IDXPROPERTIES';
KW_IF : 'IF';
KW_IGNORE : 'IGNORE';
@ -174,7 +180,8 @@ KW_INPATH : 'INPATH';
KW_INPUTDRIVER : 'INPUTDRIVER';
KW_INPUTFORMAT : 'INPUTFORMAT';
KW_INSERT : 'INSERT';
KW_INT : 'INT' 'EGER'?;
KW_INT : 'INT';
KW_INTEGER : 'INTEGER';
KW_INTERSECT : 'INTERSECT';
KW_INTERVAL : 'INTERVAL';
KW_INTO : 'INTO';
@ -216,15 +223,17 @@ KW_MATERIALIZED : 'MATERIALIZED';
KW_MERGE : 'MERGE';
KW_METADATA : 'METADATA';
KW_MINUS : 'MINUS';
KW_MINUTE : 'MINUTE' 'S'?;
KW_MONTH : 'MONTH' 'S'?;
KW_MINUTE : 'MINUTE';
KW_MINUTES : 'MINUTES';
KW_MONTH : 'MONTH';
KW_MONTHS : 'MONTHS';
KW_MORE : 'MORE';
KW_MOVE : 'MOVE';
KW_MSCK : 'MSCK';
KW_NONE : 'NONE';
KW_NORELY : 'NORELY';
KW_NOSCAN : 'NOSCAN';
KW_NOT : 'NOT' | '!';
KW_NOT : 'NOT';
KW_NOVALIDATE : 'NOVALIDATE';
KW_NO_DROP : 'NO_DROP';
KW_NULL : 'NULL';
@ -305,7 +314,8 @@ KW_SCHEDULED : 'SCHEDULED';
KW_SCHEDULING_POLICY : 'SCHEDULING_POLICY';
KW_SCHEMA : 'SCHEMA';
KW_SCHEMAS : 'SCHEMAS';
KW_SECOND : 'SECOND' 'S'?;
KW_SECOND : 'SECOND';
KW_SECONDS : 'SECONDS';
KW_SELECT : 'SELECT';
KW_SEMI : 'SEMI';
KW_SERDE : 'SERDE';
@ -388,7 +398,8 @@ KW_VECTORIZATION : 'VECTORIZATION';
KW_VIEW : 'VIEW';
KW_VIEWS : 'VIEWS';
KW_WAIT : 'WAIT';
KW_WEEK : 'WEEK' 'S'?;
KW_WEEK : 'WEEK';
KW_WEEKS : 'WEEKS';
KW_WHEN : 'WHEN';
KW_WHERE : 'WHERE';
KW_WHILE : 'WHILE';
@ -398,7 +409,8 @@ KW_WITHIN : 'WITHIN';
KW_WORK : 'WORK';
KW_WORKLOAD : 'WORKLOAD';
KW_WRITE : 'WRITE';
KW_YEAR : 'YEAR' 'S'?;
KW_YEAR : 'YEAR';
KW_YEARS : 'YEARS';
KW_ZONE : 'ZONE';
// Operators
@ -430,6 +442,7 @@ STAR : '*';
MOD : '%';
DIV : 'DIV';
BITWISENOT: '!';
AMPERSAND : '&';
TILDE : '~';
BITWISEOR : '|';
@ -525,7 +538,7 @@ RegexComponent
: 'A'..'Z' | '0'..'9' | '_'
| PLUS | STAR | QUESTION | MINUS | DOT
| LPAREN | RPAREN | LSQUARE | RSQUARE | LCURLY | RCURLY
| BITWISEXOR | BITWISEOR | DOLLAR | '!'
| BITWISEXOR | BITWISEOR | DOLLAR | BITWISENOT
;
CharSetName

116
src/grammar/hive/HiveSqlParser.g4
View File

@ -205,11 +205,11 @@ force
;
rewriteEnabled
: KW_ENABLE KW_REWRITE
: enable KW_REWRITE
;
rewriteDisabled
: KW_DISABLE KW_REWRITE
: disable KW_REWRITE
;
storedAsDirs
@ -681,7 +681,7 @@ columnNameTransformConstraint
partitionTransformType
: columnName
| (KW_YEAR | KW_MONTH | KW_DAY | KW_HOUR) LPAREN columnName RPAREN
| (year | month | day | hour) LPAREN columnName RPAREN
| (KW_TRUNCATE | KW_BUCKET) LPAREN value=Number COMMA columnName RPAREN
;
@ -822,8 +822,8 @@ enableValidateSpecification
;
enableSpecification
: KW_ENABLE
| KW_DISABLE
: enable
| disable
;
validateSpecification
@ -1021,6 +1021,7 @@ primitiveType
: KW_TINYINT
| KW_SMALLINT
| KW_INT
| KW_INTEGER
| KW_BIGINT
| KW_BOOLEAN
| KW_FLOAT
@ -1038,7 +1039,7 @@ primitiveType
//| KW_INTERVAL KW_DAY KW_TO KW_SECOND
| KW_STRING
| KW_BINARY
| KW_DECIMAL (LPAREN prec=Number (COMMA scale=Number)? RPAREN)?
| decimal (LPAREN prec=Number (COMMA scale=Number)? RPAREN)?
| (KW_VARCHAR | KW_CHAR) LPAREN length=Number RPAREN
;
@ -2170,14 +2171,14 @@ floorExpression
;
floorDateQualifiers
: KW_YEAR
: year
| KW_QUARTER
| KW_MONTH
| KW_WEEK
| KW_DAY
| KW_HOUR
| KW_MINUTE
| KW_SECOND
| month
| week
| day
| hour
| minute
| second
;
extractExpression
@ -2185,14 +2186,14 @@ extractExpression
;
timeQualifiers
: KW_YEAR
: year
| KW_QUARTER
| KW_MONTH
| KW_WEEK
| KW_DAY
| KW_HOUR
| KW_MINUTE
| KW_SECOND
| month
| week
| day
| hour
| minute
| second
;
constant
@ -2256,14 +2257,14 @@ intervalExpression
;
intervalQualifiers
: KW_YEAR KW_TO KW_MONTH
| KW_DAY KW_TO KW_SECOND
| KW_YEAR
| KW_MONTH
| KW_DAY
| KW_HOUR
| KW_MINUTE
| KW_SECOND
: year KW_TO month
| day KW_TO second
| year
| month
| day
| hour
| minute
| second
;
expression
@ -2292,6 +2293,7 @@ precedenceUnaryOperator
: PLUS
| MINUS
| TILDE
| BITWISENOT
;
precedenceUnaryPrefixExpression
@ -2524,6 +2526,7 @@ sysFuncNames
| KW_TINYINT
| KW_SMALLINT
| KW_INT
| KW_INTEGER
| KW_BIGINT
| KW_FLOAT
| KW_REAL
@ -2629,6 +2632,7 @@ nonReserved
| KW_DATABASES
| KW_DATETIME
| KW_DAY
| KW_DAYS
| KW_DAYOFWEEK
| KW_DBPROPERTIES
| KW_DCPROPERTIES
@ -2643,12 +2647,14 @@ nonReserved
| KW_DIRECTORIES
| KW_DIRECTORY
| KW_DISABLE
| KW_DISABLED
| KW_DISTRIBUTE
| KW_DISTRIBUTED
| KW_DO
| KW_DUMP
| KW_ELEM_TYPE
| KW_ENABLE
| KW_ENABLED
| KW_ENFORCED
| KW_ESCAPED
| KW_EVERY
@ -2668,6 +2674,7 @@ nonReserved
| KW_FUNCTIONS
| KW_HOLD_DDLTIME
| KW_HOUR
| KW_HOURS
| KW_IDXPROPERTIES
| KW_IGNORE
| KW_INDEX
@ -2702,7 +2709,9 @@ nonReserved
| KW_MATERIALIZED
| KW_METADATA
| KW_MINUTE
| KW_MINUTES
| KW_MONTH
| KW_MONTHS
| KW_MOVE
| KW_MSCK
| KW_NORELY
@ -2756,6 +2765,7 @@ nonReserved
| KW_SCHEMA
| KW_SCHEMAS
| KW_SECOND
| KW_SECONDS
| KW_SEMI
| KW_SERDE
| KW_SERDEPROPERTIES
@ -2812,12 +2822,14 @@ nonReserved
| KW_VIEWS
| KW_WAIT
| KW_WEEK
| KW_WEEKS
| KW_WHILE
| KW_WITHIN
| KW_WORK
| KW_WORKLOAD
| KW_WRITE
| KW_YEAR
| KW_YEARS
| KW_ZONE
;
@ -2837,6 +2849,7 @@ sql11ReservedKeywordsUsedAsFunctionName
| KW_GROUPING
| KW_IF
| KW_INT
| KW_INTEGER
| KW_MAP
| KW_REAL
| KW_SMALLINT
@ -2937,20 +2950,63 @@ activate
enable
: KW_ENABLE
| KW_ENABLED
;
disable
: KW_DISABLE
| KW_DISABLED
;
unmanaged
: KW_UNMANAGED
;
year
: KW_YEAR
| KW_YEARS
;
month
: KW_MONTH
| KW_MONTHS
;
week
: KW_WEEK
| KW_WEEKS
;
day
: KW_DAY
| KW_DAYS
;
hour
: KW_HOUR
| KW_HOURS
;
minute
: KW_MINUTE
| KW_MINUTES
;
second
: KW_SECOND
| KW_SECONDS
;
decimal
: KW_DEC
| KW_DECIMAL
| KW_NUMERIC
;
alterResourcePlanStatement
: KW_ALTER KW_RESOURCE KW_PLAN name=id_ (
KW_VALIDATE
| KW_DISABLE
| disable
| KW_SET rpAssignList
| KW_UNSET rpUnassignList
| KW_RENAME KW_TO newName=id_
@ -2962,7 +3018,7 @@ alterResourcePlanStatement
/** It might make sense to make this more generic, if something else could be enabled/disabled.
For now, it's only used for WM. Translate into another form of an alter statement. */
globalWmStatement
: (KW_ENABLE | KW_DISABLE) KW_WORKLOAD KW_MANAGEMENT
: (enable | disable) KW_WORKLOAD KW_MANAGEMENT
;
replaceResourcePlanStatement