Edit on GitHub

sqlglot.dialects.duckdb

View Source

  1from __future__ import annotations
  2
  3import typing as t
  4
  5from sqlglot import exp, generator, parser, tokens
  6from sqlglot.dialects.dialect import (
  7    Dialect,
  8    approx_count_distinct_sql,
  9    arrow_json_extract_scalar_sql,
 10    arrow_json_extract_sql,
 11    datestrtodate_sql,
 12    format_time_lambda,
 13    no_comment_column_constraint_sql,
 14    no_properties_sql,
 15    no_safe_divide_sql,
 16    pivot_column_names,
 17    rename_func,
 18    str_position_sql,
 19    str_to_time_sql,
 20    timestamptrunc_sql,
 21    timestrtotime_sql,
 22    ts_or_ds_to_date_sql,
 23)
 24from sqlglot.helper import seq_get
 25from sqlglot.tokens import TokenType
 26
 27
 28def _ts_or_ds_add_sql(self: generator.Generator, expression: exp.TsOrDsAdd) -> str:
 29    this = self.sql(expression, "this")
 30    unit = self.sql(expression, "unit").strip("'") or "DAY"
 31    return f"CAST({this} AS DATE) + {self.sql(exp.Interval(this=expression.expression, unit=unit))}"
 32
 33
 34def _date_delta_sql(self: generator.Generator, expression: exp.DateAdd | exp.DateSub) -> str:
 35    this = self.sql(expression, "this")
 36    unit = self.sql(expression, "unit").strip("'") or "DAY"
 37    op = "+" if isinstance(expression, exp.DateAdd) else "-"
 38    return f"{this} {op} {self.sql(exp.Interval(this=expression.expression, unit=unit))}"
 39
 40
 41def _array_sort_sql(self: generator.Generator, expression: exp.ArraySort) -> str:
 42    if expression.expression:
 43        self.unsupported("DUCKDB ARRAY_SORT does not support a comparator")
 44    return f"ARRAY_SORT({self.sql(expression, 'this')})"
 45
 46
 47def _sort_array_sql(self: generator.Generator, expression: exp.SortArray) -> str:
 48    this = self.sql(expression, "this")
 49    if expression.args.get("asc") == exp.false():
 50        return f"ARRAY_REVERSE_SORT({this})"
 51    return f"ARRAY_SORT({this})"
 52
 53
 54def _sort_array_reverse(args: t.List) -> exp.Expression:
 55    return exp.SortArray(this=seq_get(args, 0), asc=exp.false())
 56
 57
 58def _parse_date_diff(args: t.List) -> exp.Expression:
 59    return exp.DateDiff(this=seq_get(args, 2), expression=seq_get(args, 1), unit=seq_get(args, 0))
 60
 61
 62def _struct_sql(self: generator.Generator, expression: exp.Struct) -> str:
 63    args = [
 64        f"'{e.name or e.this.name}': {self.sql(e, 'expression')}" for e in expression.expressions
 65    ]
 66    return f"{{{', '.join(args)}}}"
 67
 68
 69def _datatype_sql(self: generator.Generator, expression: exp.DataType) -> str:
 70    if expression.is_type("array"):
 71        return f"{self.expressions(expression, flat=True)}[]"
 72    return self.datatype_sql(expression)
 73
 74
 75def _regexp_extract_sql(self: generator.Generator, expression: exp.RegexpExtract) -> str:
 76    bad_args = list(filter(expression.args.get, ("position", "occurrence")))
 77    if bad_args:
 78        self.unsupported(f"REGEXP_EXTRACT does not support arg(s) {bad_args}")
 79
 80    return self.func(
 81        "REGEXP_EXTRACT",
 82        expression.args.get("this"),
 83        expression.args.get("expression"),
 84        expression.args.get("group"),
 85    )
 86
 87
 88class DuckDB(Dialect):
 89    NULL_ORDERING = "nulls_are_last"
 90
 91    # https://duckdb.org/docs/sql/introduction.html#creating-a-new-table
 92    RESOLVES_IDENTIFIERS_AS_UPPERCASE = None
 93
 94    class Tokenizer(tokens.Tokenizer):
 95        KEYWORDS = {
 96            **tokens.Tokenizer.KEYWORDS,
 97            "~": TokenType.RLIKE,
 98            ":=": TokenType.EQ,
 99            "//": TokenType.DIV,
100            "ATTACH": TokenType.COMMAND,
101            "BINARY": TokenType.VARBINARY,
102            "BPCHAR": TokenType.TEXT,
103            "BITSTRING": TokenType.BIT,
104            "CHAR": TokenType.TEXT,
105            "CHARACTER VARYING": TokenType.TEXT,
106            "EXCLUDE": TokenType.EXCEPT,
107            "INT1": TokenType.TINYINT,
108            "LOGICAL": TokenType.BOOLEAN,
109            "NUMERIC": TokenType.DOUBLE,
110            "PIVOT_WIDER": TokenType.PIVOT,
111            "SIGNED": TokenType.INT,
112            "STRING": TokenType.VARCHAR,
113            "UBIGINT": TokenType.UBIGINT,
114            "UINTEGER": TokenType.UINT,
115            "USMALLINT": TokenType.USMALLINT,
116            "UTINYINT": TokenType.UTINYINT,
117        }
118
119    class Parser(parser.Parser):
120        CONCAT_NULL_OUTPUTS_STRING = True
121
122        FUNCTIONS = {
123            **parser.Parser.FUNCTIONS,
124            "ARRAY_LENGTH": exp.ArraySize.from_arg_list,
125            "ARRAY_SORT": exp.SortArray.from_arg_list,
126            "ARRAY_REVERSE_SORT": _sort_array_reverse,
127            "DATEDIFF": _parse_date_diff,
128            "DATE_DIFF": _parse_date_diff,
129            "EPOCH": exp.TimeToUnix.from_arg_list,
130            "EPOCH_MS": lambda args: exp.UnixToTime(
131                this=exp.Div(this=seq_get(args, 0), expression=exp.Literal.number(1000))
132            ),
133            "LIST_REVERSE_SORT": _sort_array_reverse,
134            "LIST_SORT": exp.SortArray.from_arg_list,
135            "LIST_VALUE": exp.Array.from_arg_list,
136            "REGEXP_MATCHES": exp.RegexpLike.from_arg_list,
137            "STRFTIME": format_time_lambda(exp.TimeToStr, "duckdb"),
138            "STRING_SPLIT": exp.Split.from_arg_list,
139            "STRING_SPLIT_REGEX": exp.RegexpSplit.from_arg_list,
140            "STRING_TO_ARRAY": exp.Split.from_arg_list,
141            "STRPTIME": format_time_lambda(exp.StrToTime, "duckdb"),
142            "STRUCT_PACK": exp.Struct.from_arg_list,
143            "STR_SPLIT": exp.Split.from_arg_list,
144            "STR_SPLIT_REGEX": exp.RegexpSplit.from_arg_list,
145            "TO_TIMESTAMP": exp.UnixToTime.from_arg_list,
146            "UNNEST": exp.Explode.from_arg_list,
147        }
148
149        TYPE_TOKENS = {
150            *parser.Parser.TYPE_TOKENS,
151            TokenType.UBIGINT,
152            TokenType.UINT,
153            TokenType.USMALLINT,
154            TokenType.UTINYINT,
155        }
156
157        def _pivot_column_names(self, aggregations: t.List[exp.Expression]) -> t.List[str]:
158            if len(aggregations) == 1:
159                return super()._pivot_column_names(aggregations)
160            return pivot_column_names(aggregations, dialect="duckdb")
161
162    class Generator(generator.Generator):
163        JOIN_HINTS = False
164        TABLE_HINTS = False
165        LIMIT_FETCH = "LIMIT"
166        STRUCT_DELIMITER = ("(", ")")
167        RENAME_TABLE_WITH_DB = False
168
169        TRANSFORMS = {
170            **generator.Generator.TRANSFORMS,
171            exp.ApproxDistinct: approx_count_distinct_sql,
172            exp.Array: lambda self, e: self.func("ARRAY", e.expressions[0])
173            if e.expressions and e.expressions[0].find(exp.Select)
174            else rename_func("LIST_VALUE")(self, e),
175            exp.ArraySize: rename_func("ARRAY_LENGTH"),
176            exp.ArraySort: _array_sort_sql,
177            exp.ArraySum: rename_func("LIST_SUM"),
178            exp.CommentColumnConstraint: no_comment_column_constraint_sql,
179            exp.CurrentDate: lambda self, e: "CURRENT_DATE",
180            exp.CurrentTime: lambda self, e: "CURRENT_TIME",
181            exp.CurrentTimestamp: lambda self, e: "CURRENT_TIMESTAMP",
182            exp.DayOfMonth: rename_func("DAYOFMONTH"),
183            exp.DayOfWeek: rename_func("DAYOFWEEK"),
184            exp.DayOfYear: rename_func("DAYOFYEAR"),
185            exp.DataType: _datatype_sql,
186            exp.DateAdd: _date_delta_sql,
187            exp.DateSub: _date_delta_sql,
188            exp.DateDiff: lambda self, e: self.func(
189                "DATE_DIFF", f"'{e.args.get('unit', 'day')}'", e.expression, e.this
190            ),
191            exp.DateStrToDate: datestrtodate_sql,
192            exp.DateToDi: lambda self, e: f"CAST(STRFTIME({self.sql(e, 'this')}, {DuckDB.DATEINT_FORMAT}) AS INT)",
193            exp.DiToDate: lambda self, e: f"CAST(STRPTIME(CAST({self.sql(e, 'this')} AS TEXT), {DuckDB.DATEINT_FORMAT}) AS DATE)",
194            exp.Explode: rename_func("UNNEST"),
195            exp.IntDiv: lambda self, e: self.binary(e, "//"),
196            exp.JSONExtract: arrow_json_extract_sql,
197            exp.JSONExtractScalar: arrow_json_extract_scalar_sql,
198            exp.JSONBExtract: arrow_json_extract_sql,
199            exp.JSONBExtractScalar: arrow_json_extract_scalar_sql,
200            exp.LogicalOr: rename_func("BOOL_OR"),
201            exp.LogicalAnd: rename_func("BOOL_AND"),
202            exp.Properties: no_properties_sql,
203            exp.RegexpExtract: _regexp_extract_sql,
204            exp.RegexpLike: rename_func("REGEXP_MATCHES"),
205            exp.RegexpSplit: rename_func("STR_SPLIT_REGEX"),
206            exp.SafeDivide: no_safe_divide_sql,
207            exp.Split: rename_func("STR_SPLIT"),
208            exp.SortArray: _sort_array_sql,
209            exp.StrPosition: str_position_sql,
210            exp.StrToDate: lambda self, e: f"CAST({str_to_time_sql(self, e)} AS DATE)",
211            exp.StrToTime: str_to_time_sql,
212            exp.StrToUnix: lambda self, e: f"EPOCH(STRPTIME({self.sql(e, 'this')}, {self.format_time(e)}))",
213            exp.Struct: _struct_sql,
214            exp.TimestampTrunc: timestamptrunc_sql,
215            exp.TimeStrToDate: lambda self, e: f"CAST({self.sql(e, 'this')} AS DATE)",
216            exp.TimeStrToTime: timestrtotime_sql,
217            exp.TimeStrToUnix: lambda self, e: f"EPOCH(CAST({self.sql(e, 'this')} AS TIMESTAMP))",
218            exp.TimeToStr: lambda self, e: f"STRFTIME({self.sql(e, 'this')}, {self.format_time(e)})",
219            exp.TimeToUnix: rename_func("EPOCH"),
220            exp.TsOrDiToDi: lambda self, e: f"CAST(SUBSTR(REPLACE(CAST({self.sql(e, 'this')} AS TEXT), '-', ''), 1, 8) AS INT)",
221            exp.TsOrDsAdd: _ts_or_ds_add_sql,
222            exp.TsOrDsToDate: ts_or_ds_to_date_sql("duckdb"),
223            exp.UnixToStr: lambda self, e: f"STRFTIME(TO_TIMESTAMP({self.sql(e, 'this')}), {self.format_time(e)})",
224            exp.UnixToTime: rename_func("TO_TIMESTAMP"),
225            exp.UnixToTimeStr: lambda self, e: f"CAST(TO_TIMESTAMP({self.sql(e, 'this')}) AS TEXT)",
226            exp.WeekOfYear: rename_func("WEEKOFYEAR"),
227        }
228
229        TYPE_MAPPING = {
230            **generator.Generator.TYPE_MAPPING,
231            exp.DataType.Type.BINARY: "BLOB",
232            exp.DataType.Type.CHAR: "TEXT",
233            exp.DataType.Type.FLOAT: "REAL",
234            exp.DataType.Type.NCHAR: "TEXT",
235            exp.DataType.Type.NVARCHAR: "TEXT",
236            exp.DataType.Type.UINT: "UINTEGER",
237            exp.DataType.Type.VARBINARY: "BLOB",
238            exp.DataType.Type.VARCHAR: "TEXT",
239        }
240
241        STAR_MAPPING = {**generator.Generator.STAR_MAPPING, "except": "EXCLUDE"}
242
243        UNWRAPPED_INTERVAL_VALUES = (exp.Column, exp.Literal, exp.Paren)
244
245        PROPERTIES_LOCATION = {
246            **generator.Generator.PROPERTIES_LOCATION,
247            exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED,
248        }
249
250        def interval_sql(self, expression: exp.Interval) -> str:
251            multiplier: t.Optional[int] = None
252            unit = expression.text("unit").lower()
253
254            if unit.startswith("week"):
255                multiplier = 7
256            if unit.startswith("quarter"):
257                multiplier = 90
258
259            if multiplier:
260                return f"({multiplier} * {super().interval_sql(exp.Interval(this=expression.this, unit=exp.var('day')))})"
261
262            return super().interval_sql(expression)
263
264        def tablesample_sql(
265            self, expression: exp.TableSample, seed_prefix: str = "SEED", sep: str = " AS "
266        ) -> str:
267            return super().tablesample_sql(expression, seed_prefix="REPEATABLE", sep=sep)

class DuckDB(sqlglot.dialects.dialect.Dialect): View Source

 89class DuckDB(Dialect):
 90    NULL_ORDERING = "nulls_are_last"
 91
 92    # https://duckdb.org/docs/sql/introduction.html#creating-a-new-table
 93    RESOLVES_IDENTIFIERS_AS_UPPERCASE = None
 94
 95    class Tokenizer(tokens.Tokenizer):
 96        KEYWORDS = {
 97            **tokens.Tokenizer.KEYWORDS,
 98            "~": TokenType.RLIKE,
 99            ":=": TokenType.EQ,
100            "//": TokenType.DIV,
101            "ATTACH": TokenType.COMMAND,
102            "BINARY": TokenType.VARBINARY,
103            "BPCHAR": TokenType.TEXT,
104            "BITSTRING": TokenType.BIT,
105            "CHAR": TokenType.TEXT,
106            "CHARACTER VARYING": TokenType.TEXT,
107            "EXCLUDE": TokenType.EXCEPT,
108            "INT1": TokenType.TINYINT,
109            "LOGICAL": TokenType.BOOLEAN,
110            "NUMERIC": TokenType.DOUBLE,
111            "PIVOT_WIDER": TokenType.PIVOT,
112            "SIGNED": TokenType.INT,
113            "STRING": TokenType.VARCHAR,
114            "UBIGINT": TokenType.UBIGINT,
115            "UINTEGER": TokenType.UINT,
116            "USMALLINT": TokenType.USMALLINT,
117            "UTINYINT": TokenType.UTINYINT,
118        }
119
120    class Parser(parser.Parser):
121        CONCAT_NULL_OUTPUTS_STRING = True
122
123        FUNCTIONS = {
124            **parser.Parser.FUNCTIONS,
125            "ARRAY_LENGTH": exp.ArraySize.from_arg_list,
126            "ARRAY_SORT": exp.SortArray.from_arg_list,
127            "ARRAY_REVERSE_SORT": _sort_array_reverse,
128            "DATEDIFF": _parse_date_diff,
129            "DATE_DIFF": _parse_date_diff,
130            "EPOCH": exp.TimeToUnix.from_arg_list,
131            "EPOCH_MS": lambda args: exp.UnixToTime(
132                this=exp.Div(this=seq_get(args, 0), expression=exp.Literal.number(1000))
133            ),
134            "LIST_REVERSE_SORT": _sort_array_reverse,
135            "LIST_SORT": exp.SortArray.from_arg_list,
136            "LIST_VALUE": exp.Array.from_arg_list,
137            "REGEXP_MATCHES": exp.RegexpLike.from_arg_list,
138            "STRFTIME": format_time_lambda(exp.TimeToStr, "duckdb"),
139            "STRING_SPLIT": exp.Split.from_arg_list,
140            "STRING_SPLIT_REGEX": exp.RegexpSplit.from_arg_list,
141            "STRING_TO_ARRAY": exp.Split.from_arg_list,
142            "STRPTIME": format_time_lambda(exp.StrToTime, "duckdb"),
143            "STRUCT_PACK": exp.Struct.from_arg_list,
144            "STR_SPLIT": exp.Split.from_arg_list,
145            "STR_SPLIT_REGEX": exp.RegexpSplit.from_arg_list,
146            "TO_TIMESTAMP": exp.UnixToTime.from_arg_list,
147            "UNNEST": exp.Explode.from_arg_list,
148        }
149
150        TYPE_TOKENS = {
151            *parser.Parser.TYPE_TOKENS,
152            TokenType.UBIGINT,
153            TokenType.UINT,
154            TokenType.USMALLINT,
155            TokenType.UTINYINT,
156        }
157
158        def _pivot_column_names(self, aggregations: t.List[exp.Expression]) -> t.List[str]:
159            if len(aggregations) == 1:
160                return super()._pivot_column_names(aggregations)
161            return pivot_column_names(aggregations, dialect="duckdb")
162
163    class Generator(generator.Generator):
164        JOIN_HINTS = False
165        TABLE_HINTS = False
166        LIMIT_FETCH = "LIMIT"
167        STRUCT_DELIMITER = ("(", ")")
168        RENAME_TABLE_WITH_DB = False
169
170        TRANSFORMS = {
171            **generator.Generator.TRANSFORMS,
172            exp.ApproxDistinct: approx_count_distinct_sql,
173            exp.Array: lambda self, e: self.func("ARRAY", e.expressions[0])
174            if e.expressions and e.expressions[0].find(exp.Select)
175            else rename_func("LIST_VALUE")(self, e),
176            exp.ArraySize: rename_func("ARRAY_LENGTH"),
177            exp.ArraySort: _array_sort_sql,
178            exp.ArraySum: rename_func("LIST_SUM"),
179            exp.CommentColumnConstraint: no_comment_column_constraint_sql,
180            exp.CurrentDate: lambda self, e: "CURRENT_DATE",
181            exp.CurrentTime: lambda self, e: "CURRENT_TIME",
182            exp.CurrentTimestamp: lambda self, e: "CURRENT_TIMESTAMP",
183            exp.DayOfMonth: rename_func("DAYOFMONTH"),
184            exp.DayOfWeek: rename_func("DAYOFWEEK"),
185            exp.DayOfYear: rename_func("DAYOFYEAR"),
186            exp.DataType: _datatype_sql,
187            exp.DateAdd: _date_delta_sql,
188            exp.DateSub: _date_delta_sql,
189            exp.DateDiff: lambda self, e: self.func(
190                "DATE_DIFF", f"'{e.args.get('unit', 'day')}'", e.expression, e.this
191            ),
192            exp.DateStrToDate: datestrtodate_sql,
193            exp.DateToDi: lambda self, e: f"CAST(STRFTIME({self.sql(e, 'this')}, {DuckDB.DATEINT_FORMAT}) AS INT)",
194            exp.DiToDate: lambda self, e: f"CAST(STRPTIME(CAST({self.sql(e, 'this')} AS TEXT), {DuckDB.DATEINT_FORMAT}) AS DATE)",
195            exp.Explode: rename_func("UNNEST"),
196            exp.IntDiv: lambda self, e: self.binary(e, "//"),
197            exp.JSONExtract: arrow_json_extract_sql,
198            exp.JSONExtractScalar: arrow_json_extract_scalar_sql,
199            exp.JSONBExtract: arrow_json_extract_sql,
200            exp.JSONBExtractScalar: arrow_json_extract_scalar_sql,
201            exp.LogicalOr: rename_func("BOOL_OR"),
202            exp.LogicalAnd: rename_func("BOOL_AND"),
203            exp.Properties: no_properties_sql,
204            exp.RegexpExtract: _regexp_extract_sql,
205            exp.RegexpLike: rename_func("REGEXP_MATCHES"),
206            exp.RegexpSplit: rename_func("STR_SPLIT_REGEX"),
207            exp.SafeDivide: no_safe_divide_sql,
208            exp.Split: rename_func("STR_SPLIT"),
209            exp.SortArray: _sort_array_sql,
210            exp.StrPosition: str_position_sql,
211            exp.StrToDate: lambda self, e: f"CAST({str_to_time_sql(self, e)} AS DATE)",
212            exp.StrToTime: str_to_time_sql,
213            exp.StrToUnix: lambda self, e: f"EPOCH(STRPTIME({self.sql(e, 'this')}, {self.format_time(e)}))",
214            exp.Struct: _struct_sql,
215            exp.TimestampTrunc: timestamptrunc_sql,
216            exp.TimeStrToDate: lambda self, e: f"CAST({self.sql(e, 'this')} AS DATE)",
217            exp.TimeStrToTime: timestrtotime_sql,
218            exp.TimeStrToUnix: lambda self, e: f"EPOCH(CAST({self.sql(e, 'this')} AS TIMESTAMP))",
219            exp.TimeToStr: lambda self, e: f"STRFTIME({self.sql(e, 'this')}, {self.format_time(e)})",
220            exp.TimeToUnix: rename_func("EPOCH"),
221            exp.TsOrDiToDi: lambda self, e: f"CAST(SUBSTR(REPLACE(CAST({self.sql(e, 'this')} AS TEXT), '-', ''), 1, 8) AS INT)",
222            exp.TsOrDsAdd: _ts_or_ds_add_sql,
223            exp.TsOrDsToDate: ts_or_ds_to_date_sql("duckdb"),
224            exp.UnixToStr: lambda self, e: f"STRFTIME(TO_TIMESTAMP({self.sql(e, 'this')}), {self.format_time(e)})",
225            exp.UnixToTime: rename_func("TO_TIMESTAMP"),
226            exp.UnixToTimeStr: lambda self, e: f"CAST(TO_TIMESTAMP({self.sql(e, 'this')}) AS TEXT)",
227            exp.WeekOfYear: rename_func("WEEKOFYEAR"),
228        }
229
230        TYPE_MAPPING = {
231            **generator.Generator.TYPE_MAPPING,
232            exp.DataType.Type.BINARY: "BLOB",
233            exp.DataType.Type.CHAR: "TEXT",
234            exp.DataType.Type.FLOAT: "REAL",
235            exp.DataType.Type.NCHAR: "TEXT",
236            exp.DataType.Type.NVARCHAR: "TEXT",
237            exp.DataType.Type.UINT: "UINTEGER",
238            exp.DataType.Type.VARBINARY: "BLOB",
239            exp.DataType.Type.VARCHAR: "TEXT",
240        }
241
242        STAR_MAPPING = {**generator.Generator.STAR_MAPPING, "except": "EXCLUDE"}
243
244        UNWRAPPED_INTERVAL_VALUES = (exp.Column, exp.Literal, exp.Paren)
245
246        PROPERTIES_LOCATION = {
247            **generator.Generator.PROPERTIES_LOCATION,
248            exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED,
249        }
250
251        def interval_sql(self, expression: exp.Interval) -> str:
252            multiplier: t.Optional[int] = None
253            unit = expression.text("unit").lower()
254
255            if unit.startswith("week"):
256                multiplier = 7
257            if unit.startswith("quarter"):
258                multiplier = 90
259
260            if multiplier:
261                return f"({multiplier} * {super().interval_sql(exp.Interval(this=expression.this, unit=exp.var('day')))})"
262
263            return super().interval_sql(expression)
264
265        def tablesample_sql(
266            self, expression: exp.TableSample, seed_prefix: str = "SEED", sep: str = " AS "
267        ) -> str:
268            return super().tablesample_sql(expression, seed_prefix="REPEATABLE", sep=sep)

Inherited Members

sqlglot.dialects.dialect.Dialect: get_or_raise; format_time; normalize_identifier; case_sensitive; can_identify; quote_identifier; parse; parse_into; generate; transpile; tokenize; parser; generator

class DuckDB.Tokenizer(sqlglot.tokens.Tokenizer): View Source

 95    class Tokenizer(tokens.Tokenizer):
 96        KEYWORDS = {
 97            **tokens.Tokenizer.KEYWORDS,
 98            "~": TokenType.RLIKE,
 99            ":=": TokenType.EQ,
100            "//": TokenType.DIV,
101            "ATTACH": TokenType.COMMAND,
102            "BINARY": TokenType.VARBINARY,
103            "BPCHAR": TokenType.TEXT,
104            "BITSTRING": TokenType.BIT,
105            "CHAR": TokenType.TEXT,
106            "CHARACTER VARYING": TokenType.TEXT,
107            "EXCLUDE": TokenType.EXCEPT,
108            "INT1": TokenType.TINYINT,
109            "LOGICAL": TokenType.BOOLEAN,
110            "NUMERIC": TokenType.DOUBLE,
111            "PIVOT_WIDER": TokenType.PIVOT,
112            "SIGNED": TokenType.INT,
113            "STRING": TokenType.VARCHAR,
114            "UBIGINT": TokenType.UBIGINT,
115            "UINTEGER": TokenType.UINT,
116            "USMALLINT": TokenType.USMALLINT,
117            "UTINYINT": TokenType.UTINYINT,
118        }

Inherited Members

sqlglot.tokens.Tokenizer: reset; tokenize; peek

class DuckDB.Parser(sqlglot.parser.Parser): View Source

120    class Parser(parser.Parser):
121        CONCAT_NULL_OUTPUTS_STRING = True
122
123        FUNCTIONS = {
124            **parser.Parser.FUNCTIONS,
125            "ARRAY_LENGTH": exp.ArraySize.from_arg_list,
126            "ARRAY_SORT": exp.SortArray.from_arg_list,
127            "ARRAY_REVERSE_SORT": _sort_array_reverse,
128            "DATEDIFF": _parse_date_diff,
129            "DATE_DIFF": _parse_date_diff,
130            "EPOCH": exp.TimeToUnix.from_arg_list,
131            "EPOCH_MS": lambda args: exp.UnixToTime(
132                this=exp.Div(this=seq_get(args, 0), expression=exp.Literal.number(1000))
133            ),
134            "LIST_REVERSE_SORT": _sort_array_reverse,
135            "LIST_SORT": exp.SortArray.from_arg_list,
136            "LIST_VALUE": exp.Array.from_arg_list,
137            "REGEXP_MATCHES": exp.RegexpLike.from_arg_list,
138            "STRFTIME": format_time_lambda(exp.TimeToStr, "duckdb"),
139            "STRING_SPLIT": exp.Split.from_arg_list,
140            "STRING_SPLIT_REGEX": exp.RegexpSplit.from_arg_list,
141            "STRING_TO_ARRAY": exp.Split.from_arg_list,
142            "STRPTIME": format_time_lambda(exp.StrToTime, "duckdb"),
143            "STRUCT_PACK": exp.Struct.from_arg_list,
144            "STR_SPLIT": exp.Split.from_arg_list,
145            "STR_SPLIT_REGEX": exp.RegexpSplit.from_arg_list,
146            "TO_TIMESTAMP": exp.UnixToTime.from_arg_list,
147            "UNNEST": exp.Explode.from_arg_list,
148        }
149
150        TYPE_TOKENS = {
151            *parser.Parser.TYPE_TOKENS,
152            TokenType.UBIGINT,
153            TokenType.UINT,
154            TokenType.USMALLINT,
155            TokenType.UTINYINT,
156        }
157
158        def _pivot_column_names(self, aggregations: t.List[exp.Expression]) -> t.List[str]:
159            if len(aggregations) == 1:
160                return super()._pivot_column_names(aggregations)
161            return pivot_column_names(aggregations, dialect="duckdb")

Parser consumes a list of tokens produced by the Tokenizer and produces a parsed syntax tree.

Arguments:

error_level: The desired error level. Default: ErrorLevel.IMMEDIATE
error_message_context: Determines the amount of context to capture from a query string when displaying the error message (in number of characters). Default: 100
max_errors: Maximum number of error messages to include in a raised ParseError. This is only relevant if error_level is ErrorLevel.RAISE. Default: 3

Inherited Members

sqlglot.parser.Parser: Parser; reset; parse; parse_into; check_errors; raise_error; expression; validate_expression

class DuckDB.Generator(sqlglot.generator.Generator): View Source

163    class Generator(generator.Generator):
164        JOIN_HINTS = False
165        TABLE_HINTS = False
166        LIMIT_FETCH = "LIMIT"
167        STRUCT_DELIMITER = ("(", ")")
168        RENAME_TABLE_WITH_DB = False
169
170        TRANSFORMS = {
171            **generator.Generator.TRANSFORMS,
172            exp.ApproxDistinct: approx_count_distinct_sql,
173            exp.Array: lambda self, e: self.func("ARRAY", e.expressions[0])
174            if e.expressions and e.expressions[0].find(exp.Select)
175            else rename_func("LIST_VALUE")(self, e),
176            exp.ArraySize: rename_func("ARRAY_LENGTH"),
177            exp.ArraySort: _array_sort_sql,
178            exp.ArraySum: rename_func("LIST_SUM"),
179            exp.CommentColumnConstraint: no_comment_column_constraint_sql,
180            exp.CurrentDate: lambda self, e: "CURRENT_DATE",
181            exp.CurrentTime: lambda self, e: "CURRENT_TIME",
182            exp.CurrentTimestamp: lambda self, e: "CURRENT_TIMESTAMP",
183            exp.DayOfMonth: rename_func("DAYOFMONTH"),
184            exp.DayOfWeek: rename_func("DAYOFWEEK"),
185            exp.DayOfYear: rename_func("DAYOFYEAR"),
186            exp.DataType: _datatype_sql,
187            exp.DateAdd: _date_delta_sql,
188            exp.DateSub: _date_delta_sql,
189            exp.DateDiff: lambda self, e: self.func(
190                "DATE_DIFF", f"'{e.args.get('unit', 'day')}'", e.expression, e.this
191            ),
192            exp.DateStrToDate: datestrtodate_sql,
193            exp.DateToDi: lambda self, e: f"CAST(STRFTIME({self.sql(e, 'this')}, {DuckDB.DATEINT_FORMAT}) AS INT)",
194            exp.DiToDate: lambda self, e: f"CAST(STRPTIME(CAST({self.sql(e, 'this')} AS TEXT), {DuckDB.DATEINT_FORMAT}) AS DATE)",
195            exp.Explode: rename_func("UNNEST"),
196            exp.IntDiv: lambda self, e: self.binary(e, "//"),
197            exp.JSONExtract: arrow_json_extract_sql,
198            exp.JSONExtractScalar: arrow_json_extract_scalar_sql,
199            exp.JSONBExtract: arrow_json_extract_sql,
200            exp.JSONBExtractScalar: arrow_json_extract_scalar_sql,
201            exp.LogicalOr: rename_func("BOOL_OR"),
202            exp.LogicalAnd: rename_func("BOOL_AND"),
203            exp.Properties: no_properties_sql,
204            exp.RegexpExtract: _regexp_extract_sql,
205            exp.RegexpLike: rename_func("REGEXP_MATCHES"),
206            exp.RegexpSplit: rename_func("STR_SPLIT_REGEX"),
207            exp.SafeDivide: no_safe_divide_sql,
208            exp.Split: rename_func("STR_SPLIT"),
209            exp.SortArray: _sort_array_sql,
210            exp.StrPosition: str_position_sql,
211            exp.StrToDate: lambda self, e: f"CAST({str_to_time_sql(self, e)} AS DATE)",
212            exp.StrToTime: str_to_time_sql,
213            exp.StrToUnix: lambda self, e: f"EPOCH(STRPTIME({self.sql(e, 'this')}, {self.format_time(e)}))",
214            exp.Struct: _struct_sql,
215            exp.TimestampTrunc: timestamptrunc_sql,
216            exp.TimeStrToDate: lambda self, e: f"CAST({self.sql(e, 'this')} AS DATE)",
217            exp.TimeStrToTime: timestrtotime_sql,
218            exp.TimeStrToUnix: lambda self, e: f"EPOCH(CAST({self.sql(e, 'this')} AS TIMESTAMP))",
219            exp.TimeToStr: lambda self, e: f"STRFTIME({self.sql(e, 'this')}, {self.format_time(e)})",
220            exp.TimeToUnix: rename_func("EPOCH"),
221            exp.TsOrDiToDi: lambda self, e: f"CAST(SUBSTR(REPLACE(CAST({self.sql(e, 'this')} AS TEXT), '-', ''), 1, 8) AS INT)",
222            exp.TsOrDsAdd: _ts_or_ds_add_sql,
223            exp.TsOrDsToDate: ts_or_ds_to_date_sql("duckdb"),
224            exp.UnixToStr: lambda self, e: f"STRFTIME(TO_TIMESTAMP({self.sql(e, 'this')}), {self.format_time(e)})",
225            exp.UnixToTime: rename_func("TO_TIMESTAMP"),
226            exp.UnixToTimeStr: lambda self, e: f"CAST(TO_TIMESTAMP({self.sql(e, 'this')}) AS TEXT)",
227            exp.WeekOfYear: rename_func("WEEKOFYEAR"),
228        }
229
230        TYPE_MAPPING = {
231            **generator.Generator.TYPE_MAPPING,
232            exp.DataType.Type.BINARY: "BLOB",
233            exp.DataType.Type.CHAR: "TEXT",
234            exp.DataType.Type.FLOAT: "REAL",
235            exp.DataType.Type.NCHAR: "TEXT",
236            exp.DataType.Type.NVARCHAR: "TEXT",
237            exp.DataType.Type.UINT: "UINTEGER",
238            exp.DataType.Type.VARBINARY: "BLOB",
239            exp.DataType.Type.VARCHAR: "TEXT",
240        }
241
242        STAR_MAPPING = {**generator.Generator.STAR_MAPPING, "except": "EXCLUDE"}
243
244        UNWRAPPED_INTERVAL_VALUES = (exp.Column, exp.Literal, exp.Paren)
245
246        PROPERTIES_LOCATION = {
247            **generator.Generator.PROPERTIES_LOCATION,
248            exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED,
249        }
250
251        def interval_sql(self, expression: exp.Interval) -> str:
252            multiplier: t.Optional[int] = None
253            unit = expression.text("unit").lower()
254
255            if unit.startswith("week"):
256                multiplier = 7
257            if unit.startswith("quarter"):
258                multiplier = 90
259
260            if multiplier:
261                return f"({multiplier} * {super().interval_sql(exp.Interval(this=expression.this, unit=exp.var('day')))})"
262
263            return super().interval_sql(expression)
264
265        def tablesample_sql(
266            self, expression: exp.TableSample, seed_prefix: str = "SEED", sep: str = " AS "
267        ) -> str:
268            return super().tablesample_sql(expression, seed_prefix="REPEATABLE", sep=sep)

Generator converts a given syntax tree to the corresponding SQL string.

Arguments:

pretty: Whether or not to format the produced SQL string. Default: False.
identify: Determines when an identifier should be quoted. Possible values are: False (default): Never quote, except in cases where it's mandatory by the dialect. True or 'always': Always quote. 'safe': Only quote identifiers that are case insensitive.
normalize: Whether or not to normalize identifiers to lowercase. Default: False.
pad: Determines the pad size in a formatted string. Default: 2.
indent: Determines the indentation size in a formatted string. Default: 2.
normalize_functions: Whether or not to normalize all function names. Possible values are: "upper" or True (default): Convert names to uppercase. "lower": Convert names to lowercase. False: Disables function name normalization.
unsupported_level: Determines the generator's behavior when it encounters unsupported expressions. Default ErrorLevel.WARN.
max_unsupported: Maximum number of unsupported messages to include in a raised UnsupportedError. This is only relevant if unsupported_level is ErrorLevel.RAISE. Default: 3
leading_comma: Determines whether or not the comma is leading or trailing in select expressions. This is only relevant when generating in pretty mode. Default: False
max_text_width: The max number of characters in a segment before creating new lines in pretty mode. The default is on the smaller end because the length only represents a segment and not the true line length. Default: 80
comments: Whether or not to preserve comments in the output SQL code. Default: True

def interval_sql(self, expression: sqlglot.expressions.Interval) -> str: View Source

251        def interval_sql(self, expression: exp.Interval) -> str:
252            multiplier: t.Optional[int] = None
253            unit = expression.text("unit").lower()
254
255            if unit.startswith("week"):
256                multiplier = 7
257            if unit.startswith("quarter"):
258                multiplier = 90
259
260            if multiplier:
261                return f"({multiplier} * {super().interval_sql(exp.Interval(this=expression.this, unit=exp.var('day')))})"
262
263            return super().interval_sql(expression)

def tablesample_sql( self, expression: sqlglot.expressions.TableSample, seed_prefix: str = 'SEED', sep: str = ' AS ') -> str: View Source

265        def tablesample_sql(
266            self, expression: exp.TableSample, seed_prefix: str = "SEED", sep: str = " AS "
267        ) -> str:
268            return super().tablesample_sql(expression, seed_prefix="REPEATABLE", sep=sep)

@classmethod

def can_identify(text: str, identify: str | bool = 'safe') -> bool: View Source

247    @classmethod
248    def can_identify(cls, text: str, identify: str | bool = "safe") -> bool:
249        """Checks if text can be identified given an identify option.
250
251        Args:
252            text: The text to check.
253            identify:
254                "always" or `True`: Always returns true.
255                "safe": True if the identifier is case-insensitive.
256
257        Returns:
258            Whether or not the given text can be identified.
259        """
260        if identify is True or identify == "always":
261            return True
262
263        if identify == "safe":
264            return not cls.case_sensitive(text)
265
266        return False

Checks if text can be identified given an identify option.

Arguments:

text: The text to check.
identify: "always" or True: Always returns true. "safe": True if the identifier is case-insensitive.

Returns:

Whether or not the given text can be identified.

Inherited Members

sqlglot.generator.Generator: Generator; generate; unsupported; sep; seg; pad_comment; maybe_comment; wrap; no_identify; normalize_func; indent; sql; uncache_sql; cache_sql; characterset_sql; column_sql; columnposition_sql; columndef_sql; columnconstraint_sql; autoincrementcolumnconstraint_sql; compresscolumnconstraint_sql; generatedasidentitycolumnconstraint_sql; notnullcolumnconstraint_sql; primarykeycolumnconstraint_sql; uniquecolumnconstraint_sql; createable_sql; create_sql; clone_sql; describe_sql; prepend_ctes; with_sql; cte_sql; tablealias_sql; bitstring_sql; hexstring_sql; bytestring_sql; rawstring_sql; datatypesize_sql; datatype_sql; directory_sql; delete_sql; drop_sql; except_sql; except_op; fetch_sql; filter_sql; hint_sql; index_sql; identifier_sql; inputoutputformat_sql; national_sql; partition_sql; properties_sql; root_properties; properties; with_properties; locate_properties; property_sql; likeproperty_sql; fallbackproperty_sql; journalproperty_sql; freespaceproperty_sql; checksumproperty_sql; mergeblockratioproperty_sql; datablocksizeproperty_sql; blockcompressionproperty_sql; isolatedloadingproperty_sql; lockingproperty_sql; withdataproperty_sql; insert_sql; intersect_sql; intersect_op; introducer_sql; pseudotype_sql; onconflict_sql; returning_sql; rowformatdelimitedproperty_sql; table_sql; pivot_sql; tuple_sql; update_sql; values_sql; var_sql; into_sql; from_sql; group_sql; having_sql; join_sql; lambda_sql; lateral_sql; limit_sql; offset_sql; setitem_sql; set_sql; pragma_sql; lock_sql; literal_sql; escape_str; loaddata_sql; null_sql; boolean_sql; order_sql; cluster_sql; distribute_sql; sort_sql; ordered_sql; matchrecognize_sql; query_modifiers; offset_limit_modifiers; after_having_modifiers; after_limit_modifiers; select_sql; schema_sql; schema_columns_sql; star_sql; parameter_sql; sessionparameter_sql; placeholder_sql; subquery_sql; qualify_sql; union_sql; union_op; unnest_sql; where_sql; window_sql; partition_by_sql; windowspec_sql; withingroup_sql; between_sql; bracket_sql; all_sql; any_sql; exists_sql; case_sql; constraint_sql; nextvaluefor_sql; extract_sql; trim_sql; safeconcat_sql; check_sql; foreignkey_sql; primarykey_sql; if_sql; matchagainst_sql; jsonkeyvalue_sql; jsonobject_sql; openjsoncolumndef_sql; openjson_sql; in_sql; in_unnest_op; return_sql; reference_sql; anonymous_sql; paren_sql; neg_sql; not_sql; alias_sql; aliases_sql; attimezone_sql; add_sql; and_sql; connector_sql; bitwiseand_sql; bitwiseleftshift_sql; bitwisenot_sql; bitwiseor_sql; bitwiserightshift_sql; bitwisexor_sql; cast_sql; currentdate_sql; collate_sql; command_sql; comment_sql; mergetreettlaction_sql; mergetreettl_sql; transaction_sql; commit_sql; rollback_sql; altercolumn_sql; renametable_sql; altertable_sql; droppartition_sql; addconstraint_sql; distinct_sql; ignorenulls_sql; respectnulls_sql; intdiv_sql; dpipe_sql; safedpipe_sql; div_sql; overlaps_sql; distance_sql; dot_sql; eq_sql; escape_sql; glob_sql; gt_sql; gte_sql; ilike_sql; ilikeany_sql; is_sql; like_sql; likeany_sql; similarto_sql; lt_sql; lte_sql; mod_sql; mul_sql; neq_sql; nullsafeeq_sql; nullsafeneq_sql; or_sql; slice_sql; sub_sql; trycast_sql; use_sql; binary; function_fallback_sql; func; format_args; text_width; format_time; expressions; op_expressions; naked_property; set_operation; tag_sql; token_sql; userdefinedfunction_sql; joinhint_sql; kwarg_sql; when_sql; merge_sql; tochar_sql; dictproperty_sql; dictrange_sql; dictsubproperty_sql; oncluster_sql