work around for haddock refusal to parse literal comment lines with *
in first character position get rid of code_units since these are not in sql2011 implement next value for parse the nullary functions with reserved names updates to the sql2003 file
This commit is contained in:
Jake Wheat
parent
7057241974
commit
7a7f4ba7aa
|
|
@ -38,9 +38,9 @@ lexers, this isn't 100% complete at the moment and needs fixing.
|
|||
The parsing code is aggressively left factored, and try is avoided as
|
||||
much as possible. Try is avoided because:
|
||||
|
||||
* when it is overused it makes the code hard to follow
|
||||
* when it is overused it makes the parsing code harder to debug
|
||||
* it makes the parser error messages much worse
|
||||
* when it is overused it makes the code hard to follow
|
||||
* when it is overused it makes the parsing code harder to debug
|
||||
* it makes the parser error messages much worse
|
||||
|
||||
The code could be made a bit simpler with a few extra 'trys', but this
|
||||
isn't done because of the impact on the parser error
|
||||
|
|
@ -74,9 +74,9 @@ syntax.
|
|||
|
||||
There are three big areas which are tricky to left factor:
|
||||
|
||||
* typenames
|
||||
* value expressions which can start with an identifier
|
||||
* infix and suffix operators
|
||||
* typenames
|
||||
* value expressions which can start with an identifier
|
||||
* infix and suffix operators
|
||||
|
||||
=== typenames
|
||||
|
||||
|
|
@ -97,12 +97,12 @@ error messages really bad.
|
|||
|
||||
Here is a list of these nodes:
|
||||
|
||||
* identifiers
|
||||
* function application
|
||||
* aggregate application
|
||||
* window application
|
||||
* typed literal: typename 'literal string'
|
||||
* interval literal which is like the typed literal with some extras
|
||||
* identifiers
|
||||
* function application
|
||||
* aggregate application
|
||||
* window application
|
||||
* typed literal: typename 'literal string'
|
||||
* interval literal which is like the typed literal with some extras
|
||||
|
||||
There is further ambiguity e.g. with typed literals with precision,
|
||||
functions, aggregates, etc. - these are an identifier, followed by
|
||||
|
|
@ -113,12 +113,12 @@ is.
|
|||
There is also a set of nodes which start with an identifier/keyword
|
||||
but can commit since no other syntax can start the same way:
|
||||
|
||||
* case
|
||||
* cast
|
||||
* exists, unique subquery
|
||||
* array constructor
|
||||
* multiset constructor
|
||||
* all the special syntax functions: extract, position, substring,
|
||||
* case
|
||||
* cast
|
||||
* exists, unique subquery
|
||||
* array constructor
|
||||
* multiset constructor
|
||||
* all the special syntax functions: extract, position, substring,
|
||||
convert, translate, overlay, trim, etc.
|
||||
|
||||
The interval literal mentioned above is treated in this group at the
|
||||
|
|
@ -143,10 +143,10 @@ standard which is able to eliminate a number of possibilities just in
|
|||
the grammar, which this parser allows. This is done for a number of
|
||||
reasons:
|
||||
|
||||
* it makes the parser simple - less variations
|
||||
* it should allow for dialects and extensibility more easily in the
|
||||
* it makes the parser simple - less variations
|
||||
* it should allow for dialects and extensibility more easily in the
|
||||
future (e.g. new infix binary operators with custom precedence)
|
||||
* many things which are effectively checked in the grammar in the
|
||||
* many things which are effectively checked in the grammar in the
|
||||
standard, can be checked using a typechecker or other simple static
|
||||
analysis
|
||||
|
||||
|
|
@ -481,7 +481,6 @@ TODO: this code needs heavy refactoring
|
|||
> ,return Nothing]
|
||||
> return (p,x)
|
||||
> lobUnits = choice [LobCharacters <$ keyword_ "characters"
|
||||
> ,LobCodeUnits <$ keyword_ "code_units"
|
||||
> ,LobOctets <$ keyword_ "octets"]
|
||||
> -- deal with multiset and array suffixes
|
||||
> tnSuffix x =
|
||||
|
|
@ -657,6 +656,10 @@ multiset(query expr). It must be there for compatibility or something.
|
|||
> ,keyword_ "table" >>
|
||||
> MultisetQueryCtor <$> parens queryExpr]
|
||||
|
||||
> nextValueFor :: Parser ValueExpr
|
||||
> nextValueFor = keywords_ ["next","value","for"] >>
|
||||
> NextValueFor <$> names
|
||||
|
||||
=== interval
|
||||
|
||||
interval literals are a special case and we follow the grammar less
|
||||
|
|
@ -1161,6 +1164,7 @@ fragile and could at least do with some heavy explanation.
|
|||
> ,cast
|
||||
> ,arrayCtor
|
||||
> ,multisetCtor
|
||||
> ,nextValueFor
|
||||
> ,subquery
|
||||
> ,intervalLit
|
||||
> ,specialOpKs
|
||||
|
|
@ -1740,7 +1744,7 @@ means).
|
|||
> ,"class_origin"
|
||||
> ,"coalesce"
|
||||
> ,"cobol"
|
||||
> ,"code_units"
|
||||
> --,"code_units"
|
||||
> ,"collation"
|
||||
> ,"collation_catalog"
|
||||
> ,"collation_name"
|
||||
|
|
@ -2002,13 +2006,13 @@ means).
|
|||
> ,"cube"
|
||||
> ,"current"
|
||||
> --,"current_date"
|
||||
> ,"current_default_transform_group"
|
||||
> ,"current_path"
|
||||
> ,"current_role"
|
||||
> --,"current_default_transform_group"
|
||||
> --,"current_path"
|
||||
> --,"current_role"
|
||||
> ,"current_time"
|
||||
> ,"current_timestamp"
|
||||
> ,"current_transform_group_for_type"
|
||||
> ,"current_user"
|
||||
> --,"current_user"
|
||||
> ,"cursor"
|
||||
> ,"cycle"
|
||||
> ,"date"
|
||||
|
|
@ -2052,7 +2056,7 @@ means).
|
|||
> ,"global"
|
||||
> ,"grant"
|
||||
> ,"group"
|
||||
> ,"grouping"
|
||||
> --,"grouping"
|
||||
> ,"having"
|
||||
> ,"hold"
|
||||
> --,"hour"
|
||||
|
|
@ -2088,7 +2092,7 @@ means).
|
|||
> ,"method"
|
||||
> --,"minute"
|
||||
> ,"modifies"
|
||||
> ,"module"
|
||||
> --,"module"
|
||||
> --,"month"
|
||||
> ,"multiset"
|
||||
> ,"national"
|
||||
|
|
@ -2151,7 +2155,7 @@ means).
|
|||
> --,"second"
|
||||
> ,"select"
|
||||
> ,"sensitive"
|
||||
> ,"session_user"
|
||||
> --,"session_user"
|
||||
> --,"set"
|
||||
> ,"similar"
|
||||
> ,"smallint"
|
||||
|
|
@ -2167,7 +2171,7 @@ means).
|
|||
> ,"submultiset"
|
||||
> ,"symmetric"
|
||||
> ,"system"
|
||||
> ,"system_user"
|
||||
> --,"system_user"
|
||||
> ,"table"
|
||||
> ,"then"
|
||||
> ,"time"
|
||||
|
|
@ -2187,7 +2191,7 @@ means).
|
|||
> ,"unnest"
|
||||
> ,"update"
|
||||
> ,"upper"
|
||||
> ,"user"
|
||||
> --,"user"
|
||||
> ,"using"
|
||||
> --,"value"
|
||||
> ,"values"
|
||||
|
|
|
|||
|
|
@ -218,6 +218,9 @@ which have been changed to try to improve the layout of the output.
|
|||
> valueExpr (Collate v c) =
|
||||
> valueExpr v <+> text "collate" <+> names c
|
||||
|
||||
> valueExpr (NextValueFor ns) =
|
||||
> text "next value for" <+> names ns
|
||||
|
||||
|
||||
> doubleUpQuotes :: String -> String
|
||||
> doubleUpQuotes [] = []
|
||||
|
|
@ -263,7 +266,6 @@ which have been changed to try to improve the layout of the output.
|
|||
> LobG -> text "G") m
|
||||
> <+> me (\x -> case x of
|
||||
> LobCharacters -> text "CHARACTERS"
|
||||
> LobCodeUnits -> text "CODE_UNITS"
|
||||
> LobOctets -> text "OCTETS") u)
|
||||
> typeName (CharTypeName t i cs col) =
|
||||
> names t
|
||||
|
|
|
|||
|
|
@ -158,6 +158,7 @@
|
|||
> | MultisetBinOp ValueExpr CombineOp SetQuantifier ValueExpr
|
||||
> | MultisetCtor [ValueExpr]
|
||||
> | MultisetQueryCtor QueryExpr
|
||||
> | NextValueFor [Name]
|
||||
> deriving (Eq,Show,Read,Data,Typeable)
|
||||
|
||||
> -- | Represents an identifier name, which can be quoted or unquoted.
|
||||
|
|
@ -190,7 +191,6 @@ TODO: add ref and scope, any others?
|
|||
> data LobMultiplier = LobK | LobM | LobG
|
||||
> deriving (Eq,Show,Read,Data,Typeable)
|
||||
> data LobUnits = LobCharacters
|
||||
> | LobCodeUnits
|
||||
> | LobOctets
|
||||
> deriving (Eq,Show,Read,Data,Typeable)
|
||||
|
||||
|
|
|
|||
9
TODO
9
TODO
|
|
@ -1,5 +1,11 @@
|
|||
continue 2003 review and tests
|
||||
|
||||
1. start replacing the 2003 stuff with 2011
|
||||
2. create an error message document for the website
|
||||
- base of error messages but add some more variations
|
||||
3. start thinking about tests for invalid syntax
|
||||
|
||||
|
||||
touch up the expr hack as best as can
|
||||
|
||||
careful review of token parses wrt trailing delimiters/junk
|
||||
|
|
@ -56,6 +62,8 @@ rules for changing the multi keyword parsing:
|
|||
|
||||
rough SQL 2003 todo, including tests to write:
|
||||
|
||||
switch to SQL 2011
|
||||
|
||||
now:
|
||||
review the commented out reserved keyword entries and work out how to
|
||||
fix
|
||||
|
|
@ -64,7 +72,6 @@ go through almost all the predicates
|
|||
window functions missing bits, window clauses
|
||||
from: more tests, review missing
|
||||
tablesample, unnest, etc.
|
||||
aggregates: where, filter + review
|
||||
rows review
|
||||
match missing bit
|
||||
between symmetric
|
||||
|
|
|
|||
|
|
@ -1,4 +1,4 @@
|
|||
0.4.0-dev (updated to 7a847045163feb2339ab40ebe93afe2f1c9ad813)
|
||||
0.4.0-dev (updated to 705724197463cd19dd8749dfd51e2eb8f1d02b8e)
|
||||
completely remove dependency on haskell-src-exts
|
||||
improve the error messages a great deal
|
||||
fix some trailing whitespace issues in the keyword style functions,
|
||||
|
|
@ -52,6 +52,11 @@
|
|||
quote character in the identifier
|
||||
implement complete interval literals (fixed the handling of the
|
||||
interval qualifier)
|
||||
make most of the standard reserved words actually reserved (still
|
||||
some gaps)
|
||||
change the natural in join abstract syntax to match the concrete
|
||||
syntax instead of combining natural, on and using into one field
|
||||
support filter and within group for aggregates
|
||||
0.3.1 (commit 5cba9a1cac19d66166aed2876d809aef892ff59f)
|
||||
update to work with ghc 7.8.1
|
||||
0.3.0 (commit 9e75fa93650b4f1a08d94f4225a243bcc50445ae)
|
||||
|
|
|
|||
|
|
@ -28,7 +28,7 @@ library
|
|||
Language.SQL.SimpleSQL.Parser,
|
||||
Language.SQL.SimpleSQL.Syntax
|
||||
other-extensions: TupleSections
|
||||
build-depends: base >=4.6 && <4.7,
|
||||
build-depends: base >=4.6 && <4.8,
|
||||
parsec >=3.1 && <3.2,
|
||||
mtl >=2.1 && <2.2,
|
||||
pretty >= 1.1 && < 1.2
|
||||
|
|
@ -40,7 +40,7 @@ Test-Suite Tests
|
|||
type: exitcode-stdio-1.0
|
||||
main-is: RunTests.lhs
|
||||
hs-source-dirs: .,tools
|
||||
Build-Depends: base >=4.6 && <4.7,
|
||||
Build-Depends: base >=4.6 && <4.8,
|
||||
parsec >=3.1 && <3.2,
|
||||
mtl >=2.1 && <2.2,
|
||||
pretty >= 1.1 && < 1.2,
|
||||
|
|
@ -71,7 +71,7 @@ Test-Suite Tests
|
|||
executable SQLIndent
|
||||
main-is: SQLIndent.lhs
|
||||
hs-source-dirs: .,tools
|
||||
Build-Depends: base >=4.6 && <4.7,
|
||||
Build-Depends: base >=4.6 && <4.8,
|
||||
parsec >=3.1 && <3.2,
|
||||
mtl >=2.1 && <2.2,
|
||||
pretty >= 1.1 && < 1.2
|
||||
|
|
|
|||
|
|
@ -15,7 +15,8 @@ large amount of the SQL.
|
|||
|
||||
> sql2003Tests :: TestItem
|
||||
> sql2003Tests = Group "sql2003Tests"
|
||||
> [stringLiterals
|
||||
> [Group "literals" [
|
||||
> stringLiterals
|
||||
> ,nationalCharacterStringLiterals
|
||||
> ,unicodeStringLiterals
|
||||
> ,binaryStringLiterals
|
||||
|
|
@ -23,32 +24,72 @@ large amount of the SQL.
|
|||
> ,intervalLiterals
|
||||
> ,booleanLiterals
|
||||
> ,identifiers
|
||||
> ,typeNameTests
|
||||
> ],Group "value expressions"
|
||||
> [typeNameTests
|
||||
> ,parenthesizedValueExpression
|
||||
> ,someGeneralValues
|
||||
> ,targetSpecification
|
||||
> ,contextuallyTypeValueSpec
|
||||
> --,nextValueExpression
|
||||
> ,moduleColumnRef
|
||||
> ,groupingOperation
|
||||
> --,windowFunction
|
||||
> --,caseExpression
|
||||
> --,castSpecification
|
||||
> ,nextValueExpression
|
||||
> -- subtype treatment, method invoc, static m i, new spec, attrib/method ref, deref, method ref, ref res
|
||||
> ,arrayElementReference
|
||||
> ,multisetElementReference
|
||||
> --,numericValueExpression
|
||||
> ,numericValueExpression
|
||||
> --,numericValueFunction
|
||||
> --,stringValueExpression
|
||||
> --,stringValueFunction
|
||||
> --,datetimeValueExpression
|
||||
> --,datetimeValueFunction
|
||||
> --,intervalValueExpression
|
||||
> --,intervalValueFunction
|
||||
> --,booleanValueExpression
|
||||
> --arrayValueExpression
|
||||
> ,arrayValueConstructor
|
||||
> ,multisetValueExpression
|
||||
> ,multisetValueFunction
|
||||
> ,multisetValueConstructor
|
||||
> ],Group "query expressions"
|
||||
> [
|
||||
> -- rowValueConstructor
|
||||
> --,rowValueExpression
|
||||
> --,tableValueConstructor
|
||||
> --,fromClause
|
||||
> --,joinedTable
|
||||
> --,whereClause
|
||||
> ,groupbyClause
|
||||
> groupbyClause
|
||||
> --,havingClause
|
||||
> --,windowClause
|
||||
> --,querySpecification
|
||||
> --,queryExpressions
|
||||
> ,quantifiedComparisonPredicate
|
||||
> --,querySpecifications
|
||||
> --,setOperations
|
||||
> --,withExpressions
|
||||
> ],Group "predicates"
|
||||
> [--comparisonPredicate
|
||||
> --,betweenPredicate
|
||||
> --,inPredicate
|
||||
> --,likePredicate
|
||||
> --,similarPredicae
|
||||
> --,nullPredicate
|
||||
> quantifiedComparisonPredicate
|
||||
> --,existsPredicate
|
||||
> ,uniquePredicate
|
||||
> --,normalizedPredicate
|
||||
> ,matchPredicate
|
||||
> --,overlapsPredicate
|
||||
> --,distinctPredicate
|
||||
> --,memberPredicate
|
||||
> --,submultisetPredicate
|
||||
> --,setPredicate
|
||||
> ,collateClause
|
||||
> ,aggregateFunctions
|
||||
> ,sortSpecificationList
|
||||
> ]
|
||||
> ]
|
||||
|
||||
= 5 Lexical Elements
|
||||
|
||||
|
|
@ -1001,8 +1042,6 @@ create a list of type name variations:
|
|||
> ,("blob(3M)", LobTypeName [Name "blob"] 3 (Just LobM) Nothing)
|
||||
> ,("blob(4M characters) "
|
||||
> ,LobTypeName [Name "blob"] 4 (Just LobM) (Just LobCharacters))
|
||||
> ,("blob(5 code_units) "
|
||||
> ,LobTypeName [Name "blob"] 5 Nothing (Just LobCodeUnits))
|
||||
> ,("blob(6G octets) "
|
||||
> ,LobTypeName [Name "blob"] 6 (Just LobG) (Just LobOctets))
|
||||
> ,("national character large object(7K) "
|
||||
|
|
@ -1172,7 +1211,19 @@ This is used in row type names.
|
|||
| USER
|
||||
| VALUE
|
||||
|
||||
TODO: review how the special keywords are parsed and add tests for these
|
||||
|
||||
> someGeneralValues :: TestItem
|
||||
> someGeneralValues = Group "some general values" $ map (uncurry TestValueExpr) $
|
||||
> map mkIden ["CURRENT_DEFAULT_TRANSFORM_GROUP"
|
||||
> ,"CURRENT_PATH"
|
||||
> ,"CURRENT_ROLE"
|
||||
> ,"CURRENT_USER"
|
||||
> ,"SESSION_USER"
|
||||
> ,"SYSTEM_USER"
|
||||
> ,"USER"
|
||||
> ,"VALUE"]
|
||||
> where
|
||||
> mkIden nm = (nm,Iden [Name nm])
|
||||
|
||||
<simple value specification> ::=
|
||||
<literal>
|
||||
|
|
@ -1269,8 +1320,11 @@ already covered above in the identifiers and names section
|
|||
<basic identifier chain>
|
||||
| MODULE <period> <qualified identifier> <period> <column name>
|
||||
|
||||
TODO: work out the exact syntax and add
|
||||
|
||||
> moduleColumnRef :: TestItem
|
||||
> moduleColumnRef = Group "module column ref" $ map (uncurry TestValueExpr)
|
||||
> [("MODULE.something.something", Iden [Name "MODULE"
|
||||
> ,Name "something"
|
||||
> ,Name "something"])]
|
||||
|
||||
|
||||
== 6.8 <SQL parameter reference> (p190)
|
||||
|
|
@ -1304,7 +1358,19 @@ ORDER BY department, job, "Total Empl", "Average Sal";
|
|||
|
||||
TODO: de-oracle the syntax and add as test case
|
||||
|
||||
|
||||
> groupingOperation :: TestItem
|
||||
> groupingOperation = Group "grouping operation" $ map (uncurry TestQueryExpr)
|
||||
> [("SELECT SalesQuota, SUM(SalesYTD) TotalSalesYTD,\n\
|
||||
> \ GROUPING(SalesQuota) AS Grouping\n\
|
||||
> \FROM Sales.SalesPerson\n\
|
||||
> \GROUP BY ROLLUP(SalesQuota);"
|
||||
> ,makeSelect
|
||||
> {qeSelectList = [(Iden [Name "SalesQuota"],Nothing)
|
||||
> ,(App [Name "SUM"] [Iden [Name "SalesYTD"]],Just (Name "TotalSalesYTD"))
|
||||
> ,(App [Name "GROUPING"] [Iden [Name "SalesQuota"]],Just (Name "Grouping"))]
|
||||
> ,qeFrom = [TRSimple [Name "Sales",Name "SalesPerson"]]
|
||||
> ,qeGroupBy = [Rollup [SimpleGroup (Iden [Name "SalesQuota"])]]})
|
||||
> ]
|
||||
|
||||
== 6.10 <window function> (p193)
|
||||
|
||||
|
|
@ -1323,6 +1389,10 @@ TODO: de-oracle the syntax and add as test case
|
|||
|
||||
TODO: window functions
|
||||
|
||||
> windowFunctions :: TestItem
|
||||
> windowFunctions = Group "window functions" $ map (uncurry TestValueExpr)
|
||||
> [
|
||||
> ]
|
||||
|
||||
|
||||
== 6.11 <case expression> (p197)
|
||||
|
|
@ -1371,7 +1441,10 @@ TODO: window functions
|
|||
|
||||
TODO: case expressions plus the 'abbreviations'
|
||||
|
||||
|
||||
> caseExpression :: TestItem
|
||||
> caseExpression = Group "case expression" $ map (uncurry TestValueExpr)
|
||||
> [
|
||||
> ]
|
||||
|
||||
== 6.12 <cast specification> (p200)
|
||||
|
||||
|
|
@ -1391,7 +1464,7 @@ This is already covered above
|
|||
|
||||
> nextValueExpression :: TestItem
|
||||
> nextValueExpression = Group "next value expression" $ map (uncurry TestValueExpr)
|
||||
> [("next value for a.b", undefined)
|
||||
> [("next value for a.b", NextValueFor [Name "a", Name "b"])
|
||||
> ]
|
||||
|
||||
|
||||
|
|
@ -1569,13 +1642,16 @@ Specify a numeric value.
|
|||
|
||||
> numericValueExpression :: TestItem
|
||||
> numericValueExpression = Group "numeric value expression" $ map (uncurry TestValueExpr)
|
||||
> [("a + b", undefined)
|
||||
> ,("a - b", undefined)
|
||||
> ,("a * b", undefined)
|
||||
> ,("a / b", undefined)
|
||||
> ,("+a", undefined)
|
||||
> ,("-a", undefined)
|
||||
> [("a + b", binOp "+")
|
||||
> ,("a - b", binOp "-")
|
||||
> ,("a * b", binOp "*")
|
||||
> ,("a / b", binOp "/")
|
||||
> ,("+a", prefOp "+")
|
||||
> ,("-a", prefOp "-")
|
||||
> ]
|
||||
> where
|
||||
> binOp o = BinOp (Iden [Name "a"]) [Name o] (Iden [Name "b"])
|
||||
> prefOp o = PrefixOp [Name o] (Iden [Name "a"])
|
||||
|
||||
== 6.27 <numeric value function> (p242)
|
||||
|
||||
|
|
@ -1600,12 +1676,22 @@ Specify a function yielding a value of type numeric.
|
|||
<string position expression>
|
||||
| <blob position expression>
|
||||
|
||||
> numericValueFunction :: TestItem
|
||||
> numericValueFunction = Group "numeric value function" $ map (uncurry TestValueExpr)
|
||||
> [
|
||||
|
||||
|
||||
<string position expression> ::=
|
||||
POSITION <left paren> <string value expression> IN <string value expression> [ USING <char length units> ] <right paren>
|
||||
|
||||
<blob position expression> ::=
|
||||
POSITION <left paren> <blob value expression> IN <blob value expression> <right paren>
|
||||
|
||||
> ("position (a in b)",undefined)
|
||||
> ,("position (a in b using characters)",undefined)
|
||||
> ,("position (a in b using octets)",undefined)
|
||||
|
||||
|
||||
TODO: position expressions
|
||||
|
||||
<length expression> ::=
|
||||
|
|
@ -1663,6 +1749,9 @@ TODO: extract expression
|
|||
|
||||
TODO: lots more expressions above
|
||||
|
||||
> ]
|
||||
|
||||
|
||||
== 6.28 <string value expression> (p251)
|
||||
|
||||
Specify a character string value or a binary string value.
|
||||
|
|
@ -3160,9 +3249,3 @@ TODO: review sort specifications
|
|||
> qe = makeSelect
|
||||
> {qeSelectList = [(Star,Nothing)]
|
||||
> ,qeFrom = [TRSimple [Name "t"]]}
|
||||
|
||||
TODO: what happened to the collation in order by?
|
||||
Answer: sort used to be a column reference with an optional
|
||||
collate. Since it is now a value expression, the collate doesn't need
|
||||
to be mentioned here.
|
||||
|
||||
|
|
|
|||
Loading…
Reference in a new issue