Edit on GitHub

sqlglot.dialects.duckdb

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