class TdTablePartitioningLevel(BaseSegment):
"""Partitioning Level.
https://docs.teradata.com/reader/eWpPpcMoLGQcZEoyt5AjEg/e0GX8Iw16u1SCwYvc5qXzg
partition_expression or
COLUMN [[NO] AUTO COMPRESS] [[ALL BUT] column_partition] [ADD constant]
column_partition := ([COLUMN|ROW] column_name (, column_name2, ...) NO AUTOCOMPRESS
partition_expression := CASE_N, RANGE_N, EXTRACT, expression and in case of multi-level in parenthesis
"""
type = "td_partitioning_level"
match_grammar = OneOf(
Sequence(
Ref("FunctionNameSegment"),
Bracketed(Anything(optional=True)),
),
Bracketed(
Delimited(
Sequence(
Ref("FunctionNameSegment"),
Bracketed(Anything(optional=True)),
),
delimiter=Ref("CommaSegment"),
), ),
)
class FromPivotExpressionSegment(BaseSegment):
"""A PIVOT expression.
https://cloud.google.com/bigquery/docs/reference/standard-sql/query-syntax#pivot_operator
"""
type = "from_pivot_expression"
match_grammar = Sequence("PIVOT", Bracketed(Anything()))
parse_grammar = Sequence(
"PIVOT",
Bracketed(
Delimited(
Sequence(
Ref("FunctionSegment"),
Ref("AliasExpressionSegment", optional=True),
),
),
"FOR",
Ref("SingleIdentifierGrammar"),
"IN",
Bracketed(
Delimited(
Sequence(
Ref("LiteralGrammar"),
Ref("AliasExpressionSegment", optional=True),
),
)
),
),
)
class FunctionDefinitionGrammar(BaseSegment):
"""This is the body of a `CREATE FUNCTION AS` statement."""
match_grammar = Sequence(
AnyNumberOf(
Sequence(
"LANGUAGE",
# Not really a parameter, but best fit for now.
Ref("ParameterNameSegment"),
Sequence(
"OPTIONS",
Bracketed(
Delimited(
Sequence(
Ref("ParameterNameSegment"),
Ref("EqualsSegment"),
Anything(),
),
delimiter=Ref("CommaSegment"),
)),
optional=True,
),
),
# There is some syntax not implemented here,
Sequence(
"AS",
OneOf(
Ref("DoubleQuotedLiteralSegment"),
Ref("QuotedLiteralSegment"),
Bracketed(
OneOf(Ref("ExpressionSegment"),
Ref("SelectStatementSegment"))),
),
),
))
class CreateFunctionStatementSegment(BaseSegment):
"""A `CREATE FUNCTION` statement.
This version in the TSQL dialect should be a "common subset" of the
structure of the code for those dialects.
Updated to include AS after declaration of RETURNS. Might be integrated in ANSI though.
postgres: https://www.postgresql.org/docs/9.1/sql-createfunction.html
snowflake: https://docs.snowflake.com/en/sql-reference/sql/create-function.html
bigquery: https://cloud.google.com/bigquery/docs/reference/standard-sql/user-defined-functions
tsql/mssql : https://docs.microsoft.com/en-us/sql/t-sql/statements/create-function-transact-sql?view=sql-server-ver15
"""
type = "create_function_statement"
match_grammar = Sequence(
"CREATE",
Sequence("OR", "ALTER", optional=True),
"FUNCTION",
Anything(),
)
parse_grammar = Sequence(
"CREATE",
Sequence("OR", "ALTER", optional=True),
"FUNCTION",
Ref("ObjectReferenceSegment"),
Ref("FunctionParameterListGrammar"),
Sequence( # Optional function return type
"RETURNS",
Ref("DatatypeSegment"),
optional=True,
),
Ref("FunctionDefinitionGrammar"),
)
class ScriptContentSegment(BaseSegment):
"""This represents the script content.
Because the script content could be written in
LUA, PYTHON, JAVA or R there is no further verification.
"""
type = "script_content"
match_grammar = Anything()
class FunctionDefinitionGrammar(BaseSegment):
"""This is the body of a `CREATE FUNCTION AS` statement.
Adjusted from ansi as Transact SQL does not seem to have the QuotedLiteralSegmentand Language.
Futhermore the body can contain almost anything like a function with table output.
"""
type = "function_statement"
name = "function_statement"
match_grammar = Sequence("AS", Sequence(Anything()))
class ColumnDefinitionSegment(BaseSegment):
"""A column definition, e.g. for CREATE TABLE or ALTER TABLE."""
type = "column_definition"
match_grammar = Sequence(
Ref("ColumnReferenceSegment"), # Column name
Ref("DatatypeSegment"), # Column type
Bracketed(Anything(), optional=True), # For types like VARCHAR(100)
AnyNumberOf(
Ref("ColumnOptionSegment", optional=True),
# Adding Teradata specific column definitions
Ref("TdColumnOptionSegment", optional=True),
),
)
class FromBeforeExpressionSegment(BaseSegment):
"""A BEFORE expression."""
type = "from_before_expression"
match_grammar = Sequence("BEFORE", Bracketed(Anything()))
parse_grammar = Sequence(
"BEFORE",
Bracketed(
OneOf("TIMESTAMP", "OFFSET", "STATEMENT"),
Ref("ParameterAssignerSegment"),
Ref("ExpressionSegment"),
),
)
class FromPivotExpressionSegment(BaseSegment):
"""A PIVOT expression."""
type = "from_pivot_expression"
match_grammar = Sequence("PIVOT", Bracketed(Anything()))
parse_grammar = Sequence(
"PIVOT",
Bracketed(
Ref("FunctionSegment"),
"FOR",
Ref("SingleIdentifierGrammar"),
"IN",
Bracketed(Delimited(Ref("LiteralGrammar"), delimiter=Ref("CommaSegment"))),
),
)
class CreateRoleStatementSegment(BaseSegment):
"""A `CREATE ROLE` statement.
As per:
https://www.postgresql.org/docs/current/sql-createrole.html
"""
type = "create_role_statement"
match_grammar = ansi_dialect.get_segment(
"CreateRoleStatementSegment").match_grammar.copy(
insert=[
Sequence(
Ref.keyword("WITH", optional=True),
# Very permissive for now. Anything can go here.
Anything(),
)
], )
class WithinGroupClauseSegment(BaseSegment):
"""An WITHIN GROUP clause for window functions.
https://www.postgresql.org/docs/current/functions-aggregate.html.
"""
type = "withingroup_clause"
match_grammar = Sequence(
"WITHIN",
"GROUP",
Bracketed(Anything(optional=True)),
)
parse_grammar = Sequence(
"WITHIN",
"GROUP",
Bracketed(Ref("OrderByClauseSegment", optional=True)),
)
class SelectClauseModifierSegment(BaseSegment):
"""Things that come after SELECT but before the columns."""
type = "select_clause_modifier"
match_grammar = OneOf(
Sequence("DISTINCT",
Sequence("ON", Bracketed(Anything()), optional=True)),
"ALL",
)
parse_grammar = OneOf(
Sequence(
"DISTINCT",
Sequence(
"ON",
Bracketed(
Delimited(Ref("ExpressionSegment"),
delimiter=Ref("CommaSegment"))),
optional=True,
),
),
"ALL",
)
class CreateModelStatementSegment(BaseSegment):
"""A `CREATE MODEL` statement.
https://docs.aws.amazon.com/redshift/latest/dg/r_CREATE_MODEL.html
NB: order of keywords matter
"""
type = "create_model_statement"
match_grammar = Sequence(
"CREATE",
"MODEL",
Ref("ObjectReferenceSegment"),
Sequence(
"FROM",
OneOf(
Ref("QuotedLiteralSegment"),
Bracketed(Ref("SelectableGrammar")),
Ref("ObjectReferenceSegment"),
),
optional=True,
),
Sequence(
"TARGET",
Ref("ColumnReferenceSegment"),
optional=True,
),
Sequence(
"FUNCTION",
Ref("ObjectReferenceSegment"),
Bracketed(
Delimited(Ref("DatatypeSegment")),
optional=True,
),
),
Sequence(
"RETURNS",
Ref("DatatypeSegment"),
optional=True,
),
Sequence(
"SAGEMAKER",
Ref("QuotedLiteralSegment"),
optional=True,
),
Sequence(
"IAM_ROLE",
OneOf(
"DEFAULT",
Ref("QuotedLiteralSegment"),
),
),
Sequence(
"AUTO",
OneOf(
"ON",
"OFF",
),
optional=True,
),
Sequence(
"MODEL_TYPE",
OneOf(
"XGBOOST",
"MLP",
"KMEANS",
),
optional=True,
),
Sequence(
"PROBLEM_TYPE",
OneOf(
"REGRESSION",
"BINARY_CLASSIFICATION",
"MULTICLASS_CLASSIFICATION",
),
optional=True,
),
Sequence(
"OBJECTIVE",
Ref("QuotedLiteralSegment"),
optional=True,
),
Sequence(
"PREPROCESSORS",
Ref("QuotedLiteralSegment"),
optional=True,
),
Sequence(
"HYPERPARAMETERS",
"DEFAULT",
Sequence(
"EXCEPT",
Bracketed(Delimited(Anything(), ), ),
optional=True,
),
optional=True,
),
Sequence(
"SETTINGS",
Bracketed(
Sequence(
"S3_BUCKET",
Ref("QuotedLiteralSegment"),
Sequence(
"KMS_KEY_ID",
Ref("QuotedLiteralSegment"),
optional=True,
),
Sequence(
"S3_GARBAGE_COLLECT",
OneOf(
"ON",
"OFF",
),
optional=True,
),
Sequence(
"MAX_CELLS",
Ref("NumericLiteralSegment"),
optional=True,
),
Sequence(
"MAX_RUNTIME",
Ref("NumericLiteralSegment"),
optional=True,
),
), ),
optional=True,
),
)
class BasicSegment(BaseSegment):
"""A basic segment for testing parse and expand."""
type = "basic"
match_grammar = Anything()
parse_grammar = BarKeyword
