Edit on GitHub

sqlglot.dialects.duckdb

View Source

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

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

 81class DuckDB(Dialect):
 82    class Tokenizer(tokens.Tokenizer):
 83        KEYWORDS = {
 84            **tokens.Tokenizer.KEYWORDS,
 85            "~": TokenType.RLIKE,
 86            ":=": TokenType.EQ,
 87            "ATTACH": TokenType.COMMAND,
 88            "BINARY": TokenType.VARBINARY,
 89            "BPCHAR": TokenType.TEXT,
 90            "BITSTRING": TokenType.BIT,
 91            "CHAR": TokenType.TEXT,
 92            "CHARACTER VARYING": TokenType.TEXT,
 93            "EXCLUDE": TokenType.EXCEPT,
 94            "INT1": TokenType.TINYINT,
 95            "LOGICAL": TokenType.BOOLEAN,
 96            "NUMERIC": TokenType.DOUBLE,
 97            "SIGNED": TokenType.INT,
 98            "STRING": TokenType.VARCHAR,
 99            "UBIGINT": TokenType.UBIGINT,
100            "UINTEGER": TokenType.UINT,
101            "USMALLINT": TokenType.USMALLINT,
102            "UTINYINT": TokenType.UTINYINT,
103        }
104
105    class Parser(parser.Parser):
106        FUNCTIONS = {
107            **parser.Parser.FUNCTIONS,  # type: ignore
108            "APPROX_COUNT_DISTINCT": exp.ApproxDistinct.from_arg_list,
109            "ARRAY_LENGTH": exp.ArraySize.from_arg_list,
110            "ARRAY_SORT": exp.SortArray.from_arg_list,
111            "ARRAY_REVERSE_SORT": _sort_array_reverse,
112            "EPOCH": exp.TimeToUnix.from_arg_list,
113            "EPOCH_MS": lambda args: exp.UnixToTime(
114                this=exp.Div(
115                    this=seq_get(args, 0),
116                    expression=exp.Literal.number(1000),
117                )
118            ),
119            "LIST_SORT": exp.SortArray.from_arg_list,
120            "LIST_REVERSE_SORT": _sort_array_reverse,
121            "LIST_VALUE": exp.Array.from_arg_list,
122            "REGEXP_MATCHES": exp.RegexpLike.from_arg_list,
123            "STRFTIME": format_time_lambda(exp.TimeToStr, "duckdb"),
124            "STRPTIME": format_time_lambda(exp.StrToTime, "duckdb"),
125            "STR_SPLIT": exp.Split.from_arg_list,
126            "STRING_SPLIT": exp.Split.from_arg_list,
127            "STRING_TO_ARRAY": exp.Split.from_arg_list,
128            "STR_SPLIT_REGEX": exp.RegexpSplit.from_arg_list,
129            "STRING_SPLIT_REGEX": exp.RegexpSplit.from_arg_list,
130            "STRUCT_PACK": exp.Struct.from_arg_list,
131            "TO_TIMESTAMP": exp.UnixToTime.from_arg_list,
132            "UNNEST": exp.Explode.from_arg_list,
133        }
134
135        TYPE_TOKENS = {
136            *parser.Parser.TYPE_TOKENS,
137            TokenType.UBIGINT,
138            TokenType.UINT,
139            TokenType.USMALLINT,
140            TokenType.UTINYINT,
141        }
142
143    class Generator(generator.Generator):
144        JOIN_HINTS = False
145        TABLE_HINTS = False
146        STRUCT_DELIMITER = ("(", ")")
147
148        TRANSFORMS = {
149            **generator.Generator.TRANSFORMS,  # type: ignore
150            exp.ApproxDistinct: approx_count_distinct_sql,
151            exp.Array: lambda self, e: self.func("ARRAY", e.expressions[0])
152            if isinstance(seq_get(e.expressions, 0), exp.Select)
153            else rename_func("LIST_VALUE")(self, e),
154            exp.ArraySize: rename_func("ARRAY_LENGTH"),
155            exp.ArraySort: _array_sort_sql,
156            exp.ArraySum: rename_func("LIST_SUM"),
157            exp.CommentColumnConstraint: no_comment_column_constraint_sql,
158            exp.DayOfMonth: rename_func("DAYOFMONTH"),
159            exp.DayOfWeek: rename_func("DAYOFWEEK"),
160            exp.DayOfYear: rename_func("DAYOFYEAR"),
161            exp.DataType: _datatype_sql,
162            exp.DateAdd: _date_add,
163            exp.DateDiff: lambda self, e: self.func(
164                "DATE_DIFF", e.args.get("unit") or exp.Literal.string("day"), e.expression, e.this
165            ),
166            exp.DateStrToDate: datestrtodate_sql,
167            exp.DateToDi: lambda self, e: f"CAST(STRFTIME({self.sql(e, 'this')}, {DuckDB.dateint_format}) AS INT)",
168            exp.DiToDate: lambda self, e: f"CAST(STRPTIME(CAST({self.sql(e, 'this')} AS TEXT), {DuckDB.dateint_format}) AS DATE)",
169            exp.Explode: rename_func("UNNEST"),
170            exp.JSONExtract: arrow_json_extract_sql,
171            exp.JSONExtractScalar: arrow_json_extract_scalar_sql,
172            exp.JSONBExtract: arrow_json_extract_sql,
173            exp.JSONBExtractScalar: arrow_json_extract_scalar_sql,
174            exp.LogicalOr: rename_func("BOOL_OR"),
175            exp.LogicalAnd: rename_func("BOOL_AND"),
176            exp.Pivot: no_pivot_sql,
177            exp.Properties: no_properties_sql,
178            exp.RegexpExtract: _regexp_extract_sql,
179            exp.RegexpLike: rename_func("REGEXP_MATCHES"),
180            exp.RegexpSplit: rename_func("STR_SPLIT_REGEX"),
181            exp.SafeDivide: no_safe_divide_sql,
182            exp.Split: rename_func("STR_SPLIT"),
183            exp.SortArray: _sort_array_sql,
184            exp.StrPosition: str_position_sql,
185            exp.StrToDate: lambda self, e: f"CAST({str_to_time_sql(self, e)} AS DATE)",
186            exp.StrToTime: str_to_time_sql,
187            exp.StrToUnix: lambda self, e: f"EPOCH(STRPTIME({self.sql(e, 'this')}, {self.format_time(e)}))",
188            exp.Struct: _struct_sql,
189            exp.TimestampTrunc: timestamptrunc_sql,
190            exp.TimeStrToDate: lambda self, e: f"CAST({self.sql(e, 'this')} AS DATE)",
191            exp.TimeStrToTime: timestrtotime_sql,
192            exp.TimeStrToUnix: lambda self, e: f"EPOCH(CAST({self.sql(e, 'this')} AS TIMESTAMP))",
193            exp.TimeToStr: lambda self, e: f"STRFTIME({self.sql(e, 'this')}, {self.format_time(e)})",
194            exp.TimeToUnix: rename_func("EPOCH"),
195            exp.TsOrDiToDi: lambda self, e: f"CAST(SUBSTR(REPLACE(CAST({self.sql(e, 'this')} AS TEXT), '-', ''), 1, 8) AS INT)",
196            exp.TsOrDsAdd: _ts_or_ds_add,
197            exp.TsOrDsToDate: ts_or_ds_to_date_sql("duckdb"),
198            exp.UnixToStr: lambda self, e: f"STRFTIME(TO_TIMESTAMP({self.sql(e, 'this')}), {self.format_time(e)})",
199            exp.UnixToTime: rename_func("TO_TIMESTAMP"),
200            exp.UnixToTimeStr: lambda self, e: f"CAST(TO_TIMESTAMP({self.sql(e, 'this')}) AS TEXT)",
201            exp.WeekOfYear: rename_func("WEEKOFYEAR"),
202        }
203
204        TYPE_MAPPING = {
205            **generator.Generator.TYPE_MAPPING,  # type: ignore
206            exp.DataType.Type.BINARY: "BLOB",
207            exp.DataType.Type.CHAR: "TEXT",
208            exp.DataType.Type.FLOAT: "REAL",
209            exp.DataType.Type.NCHAR: "TEXT",
210            exp.DataType.Type.NVARCHAR: "TEXT",
211            exp.DataType.Type.UINT: "UINTEGER",
212            exp.DataType.Type.VARBINARY: "BLOB",
213            exp.DataType.Type.VARCHAR: "TEXT",
214        }
215
216        STAR_MAPPING = {
217            **generator.Generator.STAR_MAPPING,
218            "except": "EXCLUDE",
219        }
220
221        PROPERTIES_LOCATION = {
222            **generator.Generator.PROPERTIES_LOCATION,  # type: ignore
223            exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED,
224        }
225
226        LIMIT_FETCH = "LIMIT"
227
228        def tablesample_sql(self, expression: exp.TableSample, seed_prefix: str = "SEED") -> str:
229            return super().tablesample_sql(expression, seed_prefix="REPEATABLE")

Inherited Members

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

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

 82    class Tokenizer(tokens.Tokenizer):
 83        KEYWORDS = {
 84            **tokens.Tokenizer.KEYWORDS,
 85            "~": TokenType.RLIKE,
 86            ":=": TokenType.EQ,
 87            "ATTACH": TokenType.COMMAND,
 88            "BINARY": TokenType.VARBINARY,
 89            "BPCHAR": TokenType.TEXT,
 90            "BITSTRING": TokenType.BIT,
 91            "CHAR": TokenType.TEXT,
 92            "CHARACTER VARYING": TokenType.TEXT,
 93            "EXCLUDE": TokenType.EXCEPT,
 94            "INT1": TokenType.TINYINT,
 95            "LOGICAL": TokenType.BOOLEAN,
 96            "NUMERIC": TokenType.DOUBLE,
 97            "SIGNED": TokenType.INT,
 98            "STRING": TokenType.VARCHAR,
 99            "UBIGINT": TokenType.UBIGINT,
100            "UINTEGER": TokenType.UINT,
101            "USMALLINT": TokenType.USMALLINT,
102            "UTINYINT": TokenType.UTINYINT,
103        }

Inherited Members

sqlglot.tokens.Tokenizer: reset; tokenize

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

105    class Parser(parser.Parser):
106        FUNCTIONS = {
107            **parser.Parser.FUNCTIONS,  # type: ignore
108            "APPROX_COUNT_DISTINCT": exp.ApproxDistinct.from_arg_list,
109            "ARRAY_LENGTH": exp.ArraySize.from_arg_list,
110            "ARRAY_SORT": exp.SortArray.from_arg_list,
111            "ARRAY_REVERSE_SORT": _sort_array_reverse,
112            "EPOCH": exp.TimeToUnix.from_arg_list,
113            "EPOCH_MS": lambda args: exp.UnixToTime(
114                this=exp.Div(
115                    this=seq_get(args, 0),
116                    expression=exp.Literal.number(1000),
117                )
118            ),
119            "LIST_SORT": exp.SortArray.from_arg_list,
120            "LIST_REVERSE_SORT": _sort_array_reverse,
121            "LIST_VALUE": exp.Array.from_arg_list,
122            "REGEXP_MATCHES": exp.RegexpLike.from_arg_list,
123            "STRFTIME": format_time_lambda(exp.TimeToStr, "duckdb"),
124            "STRPTIME": format_time_lambda(exp.StrToTime, "duckdb"),
125            "STR_SPLIT": exp.Split.from_arg_list,
126            "STRING_SPLIT": exp.Split.from_arg_list,
127            "STRING_TO_ARRAY": exp.Split.from_arg_list,
128            "STR_SPLIT_REGEX": exp.RegexpSplit.from_arg_list,
129            "STRING_SPLIT_REGEX": exp.RegexpSplit.from_arg_list,
130            "STRUCT_PACK": exp.Struct.from_arg_list,
131            "TO_TIMESTAMP": exp.UnixToTime.from_arg_list,
132            "UNNEST": exp.Explode.from_arg_list,
133        }
134
135        TYPE_TOKENS = {
136            *parser.Parser.TYPE_TOKENS,
137            TokenType.UBIGINT,
138            TokenType.UINT,
139            TokenType.USMALLINT,
140            TokenType.UTINYINT,
141        }

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

Arguments:

error_level: the desired error level. Default: ErrorLevel.RAISE
error_message_context: determines the amount of context to capture from a query string when displaying the error message (in number of characters). Default: 50.
index_offset: Index offset for arrays eg ARRAY[0] vs ARRAY[1] as the head of a list. Default: 0
alias_post_tablesample: If the table alias comes after tablesample. Default: False
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
null_ordering: Indicates the default null ordering method to use if not explicitly set. Options are "nulls_are_small", "nulls_are_large", "nulls_are_last". Default: "nulls_are_small"

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

143    class Generator(generator.Generator):
144        JOIN_HINTS = False
145        TABLE_HINTS = False
146        STRUCT_DELIMITER = ("(", ")")
147
148        TRANSFORMS = {
149            **generator.Generator.TRANSFORMS,  # type: ignore
150            exp.ApproxDistinct: approx_count_distinct_sql,
151            exp.Array: lambda self, e: self.func("ARRAY", e.expressions[0])
152            if isinstance(seq_get(e.expressions, 0), exp.Select)
153            else rename_func("LIST_VALUE")(self, e),
154            exp.ArraySize: rename_func("ARRAY_LENGTH"),
155            exp.ArraySort: _array_sort_sql,
156            exp.ArraySum: rename_func("LIST_SUM"),
157            exp.CommentColumnConstraint: no_comment_column_constraint_sql,
158            exp.DayOfMonth: rename_func("DAYOFMONTH"),
159            exp.DayOfWeek: rename_func("DAYOFWEEK"),
160            exp.DayOfYear: rename_func("DAYOFYEAR"),
161            exp.DataType: _datatype_sql,
162            exp.DateAdd: _date_add,
163            exp.DateDiff: lambda self, e: self.func(
164                "DATE_DIFF", e.args.get("unit") or exp.Literal.string("day"), e.expression, e.this
165            ),
166            exp.DateStrToDate: datestrtodate_sql,
167            exp.DateToDi: lambda self, e: f"CAST(STRFTIME({self.sql(e, 'this')}, {DuckDB.dateint_format}) AS INT)",
168            exp.DiToDate: lambda self, e: f"CAST(STRPTIME(CAST({self.sql(e, 'this')} AS TEXT), {DuckDB.dateint_format}) AS DATE)",
169            exp.Explode: rename_func("UNNEST"),
170            exp.JSONExtract: arrow_json_extract_sql,
171            exp.JSONExtractScalar: arrow_json_extract_scalar_sql,
172            exp.JSONBExtract: arrow_json_extract_sql,
173            exp.JSONBExtractScalar: arrow_json_extract_scalar_sql,
174            exp.LogicalOr: rename_func("BOOL_OR"),
175            exp.LogicalAnd: rename_func("BOOL_AND"),
176            exp.Pivot: no_pivot_sql,
177            exp.Properties: no_properties_sql,
178            exp.RegexpExtract: _regexp_extract_sql,
179            exp.RegexpLike: rename_func("REGEXP_MATCHES"),
180            exp.RegexpSplit: rename_func("STR_SPLIT_REGEX"),
181            exp.SafeDivide: no_safe_divide_sql,
182            exp.Split: rename_func("STR_SPLIT"),
183            exp.SortArray: _sort_array_sql,
184            exp.StrPosition: str_position_sql,
185            exp.StrToDate: lambda self, e: f"CAST({str_to_time_sql(self, e)} AS DATE)",
186            exp.StrToTime: str_to_time_sql,
187            exp.StrToUnix: lambda self, e: f"EPOCH(STRPTIME({self.sql(e, 'this')}, {self.format_time(e)}))",
188            exp.Struct: _struct_sql,
189            exp.TimestampTrunc: timestamptrunc_sql,
190            exp.TimeStrToDate: lambda self, e: f"CAST({self.sql(e, 'this')} AS DATE)",
191            exp.TimeStrToTime: timestrtotime_sql,
192            exp.TimeStrToUnix: lambda self, e: f"EPOCH(CAST({self.sql(e, 'this')} AS TIMESTAMP))",
193            exp.TimeToStr: lambda self, e: f"STRFTIME({self.sql(e, 'this')}, {self.format_time(e)})",
194            exp.TimeToUnix: rename_func("EPOCH"),
195            exp.TsOrDiToDi: lambda self, e: f"CAST(SUBSTR(REPLACE(CAST({self.sql(e, 'this')} AS TEXT), '-', ''), 1, 8) AS INT)",
196            exp.TsOrDsAdd: _ts_or_ds_add,
197            exp.TsOrDsToDate: ts_or_ds_to_date_sql("duckdb"),
198            exp.UnixToStr: lambda self, e: f"STRFTIME(TO_TIMESTAMP({self.sql(e, 'this')}), {self.format_time(e)})",
199            exp.UnixToTime: rename_func("TO_TIMESTAMP"),
200            exp.UnixToTimeStr: lambda self, e: f"CAST(TO_TIMESTAMP({self.sql(e, 'this')}) AS TEXT)",
201            exp.WeekOfYear: rename_func("WEEKOFYEAR"),
202        }
203
204        TYPE_MAPPING = {
205            **generator.Generator.TYPE_MAPPING,  # type: ignore
206            exp.DataType.Type.BINARY: "BLOB",
207            exp.DataType.Type.CHAR: "TEXT",
208            exp.DataType.Type.FLOAT: "REAL",
209            exp.DataType.Type.NCHAR: "TEXT",
210            exp.DataType.Type.NVARCHAR: "TEXT",
211            exp.DataType.Type.UINT: "UINTEGER",
212            exp.DataType.Type.VARBINARY: "BLOB",
213            exp.DataType.Type.VARCHAR: "TEXT",
214        }
215
216        STAR_MAPPING = {
217            **generator.Generator.STAR_MAPPING,
218            "except": "EXCLUDE",
219        }
220
221        PROPERTIES_LOCATION = {
222            **generator.Generator.PROPERTIES_LOCATION,  # type: ignore
223            exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED,
224        }
225
226        LIMIT_FETCH = "LIMIT"
227
228        def tablesample_sql(self, expression: exp.TableSample, seed_prefix: str = "SEED") -> str:
229            return super().tablesample_sql(expression, seed_prefix="REPEATABLE")

Generator interprets the given syntax tree and produces a SQL string as an output.

Arguments:

time_mapping (dict): the dictionary of custom time mappings in which the key represents a python time format and the output the target time format
time_trie (trie): a trie of the time_mapping keys
pretty (bool): if set to True the returned string will be formatted. Default: False.
quote_start (str): specifies which starting character to use to delimit quotes. Default: '.
quote_end (str): specifies which ending character to use to delimit quotes. Default: '.
identifier_start (str): specifies which starting character to use to delimit identifiers. Default: ".
identifier_end (str): specifies which ending character to use to delimit identifiers. Default: ".
identify (bool | str): 'always': always quote, 'safe': quote identifiers if they don't contain an upcase, True defaults to always.
normalize (bool): if set to True all identifiers will lower cased
string_escape (str): specifies a string escape character. Default: '.
identifier_escape (str): specifies an identifier escape character. Default: ".
pad (int): determines padding in a formatted string. Default: 2.
indent (int): determines the size of indentation in a formatted string. Default: 4.
unnest_column_only (bool): if true unnest table aliases are considered only as column aliases
normalize_functions (str): normalize function names, "upper", "lower", or None Default: "upper"
alias_post_tablesample (bool): if the table alias comes after tablesample Default: False
unsupported_level (ErrorLevel): determines the generator's behavior when it encounters unsupported expressions. Default ErrorLevel.WARN.
null_ordering (str): Indicates the default null ordering method to use if not explicitly set. Options are "nulls_are_small", "nulls_are_large", "nulls_are_last". Default: "nulls_are_small"
max_unsupported (int): 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 (bool): if the the comma is leading or trailing in select statements 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') -> str: View Source

228        def tablesample_sql(self, expression: exp.TableSample, seed_prefix: str = "SEED") -> str:
229            return super().tablesample_sql(expression, seed_prefix="REPEATABLE")

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; create_sql; describe_sql; prepend_ctes; with_sql; cte_sql; tablealias_sql; bitstring_sql; hexstring_sql; bytestring_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; afterjournalproperty_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; select_sql; schema_sql; star_sql; structkwarg_sql; parameter_sql; sessionparameter_sql; placeholder_sql; subquery_sql; qualify_sql; union_sql; union_op; unnest_sql; where_sql; window_sql; partition_by_sql; window_spec_sql; withingroup_sql; between_sql; bracket_sql; all_sql; any_sql; exists_sql; case_sql; constraint_sql; extract_sql; trim_sql; concat_sql; check_sql; foreignkey_sql; primarykey_sql; unique_sql; if_sql; matchagainst_sql; jsonkeyvalue_sql; jsonobject_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; 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; 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