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_pivot_sql, 15 no_properties_sql, 16 no_safe_divide_sql, 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_add_sql(self: generator.Generator, expression: exp.DateAdd) -> str: 35 this = self.sql(expression, "this") 36 unit = self.sql(expression, "unit").strip("'") or "DAY" 37 return f"{this} + {self.sql(exp.Interval(this=expression.expression, unit=unit))}" 38 39 40def _array_sort_sql(self: generator.Generator, expression: exp.ArraySort) -> str: 41 if expression.expression: 42 self.unsupported("DUCKDB ARRAY_SORT does not support a comparator") 43 return f"ARRAY_SORT({self.sql(expression, 'this')})" 44 45 46def _sort_array_sql(self: generator.Generator, expression: exp.SortArray) -> str: 47 this = self.sql(expression, "this") 48 if expression.args.get("asc") == exp.false(): 49 return f"ARRAY_REVERSE_SORT({this})" 50 return f"ARRAY_SORT({this})" 51 52 53def _sort_array_reverse(args: t.Sequence) -> exp.Expression: 54 return exp.SortArray(this=seq_get(args, 0), asc=exp.false()) 55 56 57def _parse_date_diff(args: t.Sequence) -> exp.Expression: 58 return exp.DateDiff( 59 this=seq_get(args, 2), 60 expression=seq_get(args, 1), 61 unit=seq_get(args, 0), 62 ) 63 64 65def _struct_sql(self: generator.Generator, expression: exp.Struct) -> str: 66 args = [ 67 f"'{e.name or e.this.name}': {self.sql(e, 'expression')}" for e in expression.expressions 68 ] 69 return f"{{{', '.join(args)}}}" 70 71 72def _datatype_sql(self: generator.Generator, expression: exp.DataType) -> str: 73 if expression.this == exp.DataType.Type.ARRAY: 74 return f"{self.expressions(expression, flat=True)}[]" 75 return self.datatype_sql(expression) 76 77 78def _regexp_extract_sql(self: generator.Generator, expression: exp.RegexpExtract) -> str: 79 bad_args = list(filter(expression.args.get, ("position", "occurrence"))) 80 if bad_args: 81 self.unsupported(f"REGEXP_EXTRACT does not support arg(s) {bad_args}") 82 83 return self.func( 84 "REGEXP_EXTRACT", 85 expression.args.get("this"), 86 expression.args.get("expression"), 87 expression.args.get("group"), 88 ) 89 90 91class DuckDB(Dialect): 92 class Tokenizer(tokens.Tokenizer): 93 KEYWORDS = { 94 **tokens.Tokenizer.KEYWORDS, 95 "~": TokenType.RLIKE, 96 ":=": TokenType.EQ, 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 "SIGNED": TokenType.INT, 108 "STRING": TokenType.VARCHAR, 109 "UBIGINT": TokenType.UBIGINT, 110 "UINTEGER": TokenType.UINT, 111 "USMALLINT": TokenType.USMALLINT, 112 "UTINYINT": TokenType.UTINYINT, 113 } 114 115 class Parser(parser.Parser): 116 FUNCTIONS = { 117 **parser.Parser.FUNCTIONS, # type: ignore 118 "APPROX_COUNT_DISTINCT": exp.ApproxDistinct.from_arg_list, 119 "ARRAY_LENGTH": exp.ArraySize.from_arg_list, 120 "ARRAY_SORT": exp.SortArray.from_arg_list, 121 "ARRAY_REVERSE_SORT": _sort_array_reverse, 122 "DATEDIFF": _parse_date_diff, 123 "DATE_DIFF": _parse_date_diff, 124 "EPOCH": exp.TimeToUnix.from_arg_list, 125 "EPOCH_MS": lambda args: exp.UnixToTime( 126 this=exp.Div( 127 this=seq_get(args, 0), 128 expression=exp.Literal.number(1000), 129 ) 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 class Generator(generator.Generator): 156 JOIN_HINTS = False 157 TABLE_HINTS = False 158 LIMIT_FETCH = "LIMIT" 159 STRUCT_DELIMITER = ("(", ")") 160 161 TRANSFORMS = { 162 **generator.Generator.TRANSFORMS, 163 exp.ApproxDistinct: approx_count_distinct_sql, 164 exp.Array: lambda self, e: self.func("ARRAY", e.expressions[0]) 165 if isinstance(seq_get(e.expressions, 0), exp.Select) 166 else rename_func("LIST_VALUE")(self, e), 167 exp.ArraySize: rename_func("ARRAY_LENGTH"), 168 exp.ArraySort: _array_sort_sql, 169 exp.ArraySum: rename_func("LIST_SUM"), 170 exp.CommentColumnConstraint: no_comment_column_constraint_sql, 171 exp.CurrentDate: lambda self, e: "CURRENT_DATE", 172 exp.CurrentTime: lambda self, e: "CURRENT_TIME", 173 exp.CurrentTimestamp: lambda self, e: "CURRENT_TIMESTAMP", 174 exp.DayOfMonth: rename_func("DAYOFMONTH"), 175 exp.DayOfWeek: rename_func("DAYOFWEEK"), 176 exp.DayOfYear: rename_func("DAYOFYEAR"), 177 exp.DataType: _datatype_sql, 178 exp.DateAdd: _date_add_sql, 179 exp.DateDiff: lambda self, e: self.func( 180 "DATE_DIFF", f"'{e.args.get('unit', 'day')}'", e.expression, e.this 181 ), 182 exp.DateStrToDate: datestrtodate_sql, 183 exp.DateToDi: lambda self, e: f"CAST(STRFTIME({self.sql(e, 'this')}, {DuckDB.dateint_format}) AS INT)", 184 exp.DiToDate: lambda self, e: f"CAST(STRPTIME(CAST({self.sql(e, 'this')} AS TEXT), {DuckDB.dateint_format}) AS DATE)", 185 exp.Explode: rename_func("UNNEST"), 186 exp.JSONExtract: arrow_json_extract_sql, 187 exp.JSONExtractScalar: arrow_json_extract_scalar_sql, 188 exp.JSONBExtract: arrow_json_extract_sql, 189 exp.JSONBExtractScalar: arrow_json_extract_scalar_sql, 190 exp.LogicalOr: rename_func("BOOL_OR"), 191 exp.LogicalAnd: rename_func("BOOL_AND"), 192 exp.Pivot: no_pivot_sql, 193 exp.Properties: no_properties_sql, 194 exp.RegexpExtract: _regexp_extract_sql, 195 exp.RegexpLike: rename_func("REGEXP_MATCHES"), 196 exp.RegexpSplit: rename_func("STR_SPLIT_REGEX"), 197 exp.SafeDivide: no_safe_divide_sql, 198 exp.Split: rename_func("STR_SPLIT"), 199 exp.SortArray: _sort_array_sql, 200 exp.StrPosition: str_position_sql, 201 exp.StrToDate: lambda self, e: f"CAST({str_to_time_sql(self, e)} AS DATE)", 202 exp.StrToTime: str_to_time_sql, 203 exp.StrToUnix: lambda self, e: f"EPOCH(STRPTIME({self.sql(e, 'this')}, {self.format_time(e)}))", 204 exp.Struct: _struct_sql, 205 exp.TimestampTrunc: timestamptrunc_sql, 206 exp.TimeStrToDate: lambda self, e: f"CAST({self.sql(e, 'this')} AS DATE)", 207 exp.TimeStrToTime: timestrtotime_sql, 208 exp.TimeStrToUnix: lambda self, e: f"EPOCH(CAST({self.sql(e, 'this')} AS TIMESTAMP))", 209 exp.TimeToStr: lambda self, e: f"STRFTIME({self.sql(e, 'this')}, {self.format_time(e)})", 210 exp.TimeToUnix: rename_func("EPOCH"), 211 exp.TsOrDiToDi: lambda self, e: f"CAST(SUBSTR(REPLACE(CAST({self.sql(e, 'this')} AS TEXT), '-', ''), 1, 8) AS INT)", 212 exp.TsOrDsAdd: _ts_or_ds_add_sql, 213 exp.TsOrDsToDate: ts_or_ds_to_date_sql("duckdb"), 214 exp.UnixToStr: lambda self, e: f"STRFTIME(TO_TIMESTAMP({self.sql(e, 'this')}), {self.format_time(e)})", 215 exp.UnixToTime: rename_func("TO_TIMESTAMP"), 216 exp.UnixToTimeStr: lambda self, e: f"CAST(TO_TIMESTAMP({self.sql(e, 'this')}) AS TEXT)", 217 exp.WeekOfYear: rename_func("WEEKOFYEAR"), 218 } 219 220 TYPE_MAPPING = { 221 **generator.Generator.TYPE_MAPPING, 222 exp.DataType.Type.BINARY: "BLOB", 223 exp.DataType.Type.CHAR: "TEXT", 224 exp.DataType.Type.FLOAT: "REAL", 225 exp.DataType.Type.NCHAR: "TEXT", 226 exp.DataType.Type.NVARCHAR: "TEXT", 227 exp.DataType.Type.UINT: "UINTEGER", 228 exp.DataType.Type.VARBINARY: "BLOB", 229 exp.DataType.Type.VARCHAR: "TEXT", 230 } 231 232 STAR_MAPPING = {**generator.Generator.STAR_MAPPING, "except": "EXCLUDE"} 233 234 PROPERTIES_LOCATION = { 235 **generator.Generator.PROPERTIES_LOCATION, # type: ignore 236 exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED, 237 } 238 239 def tablesample_sql( 240 self, expression: exp.TableSample, seed_prefix: str = "SEED", sep=" AS " 241 ) -> str: 242 return super().tablesample_sql(expression, seed_prefix="REPEATABLE", sep=sep)
92class DuckDB(Dialect): 93 class Tokenizer(tokens.Tokenizer): 94 KEYWORDS = { 95 **tokens.Tokenizer.KEYWORDS, 96 "~": TokenType.RLIKE, 97 ":=": TokenType.EQ, 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 "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 FUNCTIONS = { 118 **parser.Parser.FUNCTIONS, # type: ignore 119 "APPROX_COUNT_DISTINCT": exp.ApproxDistinct.from_arg_list, 120 "ARRAY_LENGTH": exp.ArraySize.from_arg_list, 121 "ARRAY_SORT": exp.SortArray.from_arg_list, 122 "ARRAY_REVERSE_SORT": _sort_array_reverse, 123 "DATEDIFF": _parse_date_diff, 124 "DATE_DIFF": _parse_date_diff, 125 "EPOCH": exp.TimeToUnix.from_arg_list, 126 "EPOCH_MS": lambda args: exp.UnixToTime( 127 this=exp.Div( 128 this=seq_get(args, 0), 129 expression=exp.Literal.number(1000), 130 ) 131 ), 132 "LIST_REVERSE_SORT": _sort_array_reverse, 133 "LIST_SORT": exp.SortArray.from_arg_list, 134 "LIST_VALUE": exp.Array.from_arg_list, 135 "REGEXP_MATCHES": exp.RegexpLike.from_arg_list, 136 "STRFTIME": format_time_lambda(exp.TimeToStr, "duckdb"), 137 "STRING_SPLIT": exp.Split.from_arg_list, 138 "STRING_SPLIT_REGEX": exp.RegexpSplit.from_arg_list, 139 "STRING_TO_ARRAY": exp.Split.from_arg_list, 140 "STRPTIME": format_time_lambda(exp.StrToTime, "duckdb"), 141 "STRUCT_PACK": exp.Struct.from_arg_list, 142 "STR_SPLIT": exp.Split.from_arg_list, 143 "STR_SPLIT_REGEX": exp.RegexpSplit.from_arg_list, 144 "TO_TIMESTAMP": exp.UnixToTime.from_arg_list, 145 "UNNEST": exp.Explode.from_arg_list, 146 } 147 148 TYPE_TOKENS = { 149 *parser.Parser.TYPE_TOKENS, 150 TokenType.UBIGINT, 151 TokenType.UINT, 152 TokenType.USMALLINT, 153 TokenType.UTINYINT, 154 } 155 156 class Generator(generator.Generator): 157 JOIN_HINTS = False 158 TABLE_HINTS = False 159 LIMIT_FETCH = "LIMIT" 160 STRUCT_DELIMITER = ("(", ")") 161 162 TRANSFORMS = { 163 **generator.Generator.TRANSFORMS, 164 exp.ApproxDistinct: approx_count_distinct_sql, 165 exp.Array: lambda self, e: self.func("ARRAY", e.expressions[0]) 166 if isinstance(seq_get(e.expressions, 0), exp.Select) 167 else rename_func("LIST_VALUE")(self, e), 168 exp.ArraySize: rename_func("ARRAY_LENGTH"), 169 exp.ArraySort: _array_sort_sql, 170 exp.ArraySum: rename_func("LIST_SUM"), 171 exp.CommentColumnConstraint: no_comment_column_constraint_sql, 172 exp.CurrentDate: lambda self, e: "CURRENT_DATE", 173 exp.CurrentTime: lambda self, e: "CURRENT_TIME", 174 exp.CurrentTimestamp: lambda self, e: "CURRENT_TIMESTAMP", 175 exp.DayOfMonth: rename_func("DAYOFMONTH"), 176 exp.DayOfWeek: rename_func("DAYOFWEEK"), 177 exp.DayOfYear: rename_func("DAYOFYEAR"), 178 exp.DataType: _datatype_sql, 179 exp.DateAdd: _date_add_sql, 180 exp.DateDiff: lambda self, e: self.func( 181 "DATE_DIFF", f"'{e.args.get('unit', 'day')}'", e.expression, e.this 182 ), 183 exp.DateStrToDate: datestrtodate_sql, 184 exp.DateToDi: lambda self, e: f"CAST(STRFTIME({self.sql(e, 'this')}, {DuckDB.dateint_format}) AS INT)", 185 exp.DiToDate: lambda self, e: f"CAST(STRPTIME(CAST({self.sql(e, 'this')} AS TEXT), {DuckDB.dateint_format}) AS DATE)", 186 exp.Explode: rename_func("UNNEST"), 187 exp.JSONExtract: arrow_json_extract_sql, 188 exp.JSONExtractScalar: arrow_json_extract_scalar_sql, 189 exp.JSONBExtract: arrow_json_extract_sql, 190 exp.JSONBExtractScalar: arrow_json_extract_scalar_sql, 191 exp.LogicalOr: rename_func("BOOL_OR"), 192 exp.LogicalAnd: rename_func("BOOL_AND"), 193 exp.Pivot: no_pivot_sql, 194 exp.Properties: no_properties_sql, 195 exp.RegexpExtract: _regexp_extract_sql, 196 exp.RegexpLike: rename_func("REGEXP_MATCHES"), 197 exp.RegexpSplit: rename_func("STR_SPLIT_REGEX"), 198 exp.SafeDivide: no_safe_divide_sql, 199 exp.Split: rename_func("STR_SPLIT"), 200 exp.SortArray: _sort_array_sql, 201 exp.StrPosition: str_position_sql, 202 exp.StrToDate: lambda self, e: f"CAST({str_to_time_sql(self, e)} AS DATE)", 203 exp.StrToTime: str_to_time_sql, 204 exp.StrToUnix: lambda self, e: f"EPOCH(STRPTIME({self.sql(e, 'this')}, {self.format_time(e)}))", 205 exp.Struct: _struct_sql, 206 exp.TimestampTrunc: timestamptrunc_sql, 207 exp.TimeStrToDate: lambda self, e: f"CAST({self.sql(e, 'this')} AS DATE)", 208 exp.TimeStrToTime: timestrtotime_sql, 209 exp.TimeStrToUnix: lambda self, e: f"EPOCH(CAST({self.sql(e, 'this')} AS TIMESTAMP))", 210 exp.TimeToStr: lambda self, e: f"STRFTIME({self.sql(e, 'this')}, {self.format_time(e)})", 211 exp.TimeToUnix: rename_func("EPOCH"), 212 exp.TsOrDiToDi: lambda self, e: f"CAST(SUBSTR(REPLACE(CAST({self.sql(e, 'this')} AS TEXT), '-', ''), 1, 8) AS INT)", 213 exp.TsOrDsAdd: _ts_or_ds_add_sql, 214 exp.TsOrDsToDate: ts_or_ds_to_date_sql("duckdb"), 215 exp.UnixToStr: lambda self, e: f"STRFTIME(TO_TIMESTAMP({self.sql(e, 'this')}), {self.format_time(e)})", 216 exp.UnixToTime: rename_func("TO_TIMESTAMP"), 217 exp.UnixToTimeStr: lambda self, e: f"CAST(TO_TIMESTAMP({self.sql(e, 'this')}) AS TEXT)", 218 exp.WeekOfYear: rename_func("WEEKOFYEAR"), 219 } 220 221 TYPE_MAPPING = { 222 **generator.Generator.TYPE_MAPPING, 223 exp.DataType.Type.BINARY: "BLOB", 224 exp.DataType.Type.CHAR: "TEXT", 225 exp.DataType.Type.FLOAT: "REAL", 226 exp.DataType.Type.NCHAR: "TEXT", 227 exp.DataType.Type.NVARCHAR: "TEXT", 228 exp.DataType.Type.UINT: "UINTEGER", 229 exp.DataType.Type.VARBINARY: "BLOB", 230 exp.DataType.Type.VARCHAR: "TEXT", 231 } 232 233 STAR_MAPPING = {**generator.Generator.STAR_MAPPING, "except": "EXCLUDE"} 234 235 PROPERTIES_LOCATION = { 236 **generator.Generator.PROPERTIES_LOCATION, # type: ignore 237 exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED, 238 } 239 240 def tablesample_sql( 241 self, expression: exp.TableSample, seed_prefix: str = "SEED", sep=" AS " 242 ) -> str: 243 return super().tablesample_sql(expression, seed_prefix="REPEATABLE", sep=sep)
93 class Tokenizer(tokens.Tokenizer): 94 KEYWORDS = { 95 **tokens.Tokenizer.KEYWORDS, 96 "~": TokenType.RLIKE, 97 ":=": TokenType.EQ, 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 "SIGNED": TokenType.INT, 109 "STRING": TokenType.VARCHAR, 110 "UBIGINT": TokenType.UBIGINT, 111 "UINTEGER": TokenType.UINT, 112 "USMALLINT": TokenType.USMALLINT, 113 "UTINYINT": TokenType.UTINYINT, 114 }
Inherited Members
116 class Parser(parser.Parser): 117 FUNCTIONS = { 118 **parser.Parser.FUNCTIONS, # type: ignore 119 "APPROX_COUNT_DISTINCT": exp.ApproxDistinct.from_arg_list, 120 "ARRAY_LENGTH": exp.ArraySize.from_arg_list, 121 "ARRAY_SORT": exp.SortArray.from_arg_list, 122 "ARRAY_REVERSE_SORT": _sort_array_reverse, 123 "DATEDIFF": _parse_date_diff, 124 "DATE_DIFF": _parse_date_diff, 125 "EPOCH": exp.TimeToUnix.from_arg_list, 126 "EPOCH_MS": lambda args: exp.UnixToTime( 127 this=exp.Div( 128 this=seq_get(args, 0), 129 expression=exp.Literal.number(1000), 130 ) 131 ), 132 "LIST_REVERSE_SORT": _sort_array_reverse, 133 "LIST_SORT": exp.SortArray.from_arg_list, 134 "LIST_VALUE": exp.Array.from_arg_list, 135 "REGEXP_MATCHES": exp.RegexpLike.from_arg_list, 136 "STRFTIME": format_time_lambda(exp.TimeToStr, "duckdb"), 137 "STRING_SPLIT": exp.Split.from_arg_list, 138 "STRING_SPLIT_REGEX": exp.RegexpSplit.from_arg_list, 139 "STRING_TO_ARRAY": exp.Split.from_arg_list, 140 "STRPTIME": format_time_lambda(exp.StrToTime, "duckdb"), 141 "STRUCT_PACK": exp.Struct.from_arg_list, 142 "STR_SPLIT": exp.Split.from_arg_list, 143 "STR_SPLIT_REGEX": exp.RegexpSplit.from_arg_list, 144 "TO_TIMESTAMP": exp.UnixToTime.from_arg_list, 145 "UNNEST": exp.Explode.from_arg_list, 146 } 147 148 TYPE_TOKENS = { 149 *parser.Parser.TYPE_TOKENS, 150 TokenType.UBIGINT, 151 TokenType.UINT, 152 TokenType.USMALLINT, 153 TokenType.UTINYINT, 154 }
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
156 class Generator(generator.Generator): 157 JOIN_HINTS = False 158 TABLE_HINTS = False 159 LIMIT_FETCH = "LIMIT" 160 STRUCT_DELIMITER = ("(", ")") 161 162 TRANSFORMS = { 163 **generator.Generator.TRANSFORMS, 164 exp.ApproxDistinct: approx_count_distinct_sql, 165 exp.Array: lambda self, e: self.func("ARRAY", e.expressions[0]) 166 if isinstance(seq_get(e.expressions, 0), exp.Select) 167 else rename_func("LIST_VALUE")(self, e), 168 exp.ArraySize: rename_func("ARRAY_LENGTH"), 169 exp.ArraySort: _array_sort_sql, 170 exp.ArraySum: rename_func("LIST_SUM"), 171 exp.CommentColumnConstraint: no_comment_column_constraint_sql, 172 exp.CurrentDate: lambda self, e: "CURRENT_DATE", 173 exp.CurrentTime: lambda self, e: "CURRENT_TIME", 174 exp.CurrentTimestamp: lambda self, e: "CURRENT_TIMESTAMP", 175 exp.DayOfMonth: rename_func("DAYOFMONTH"), 176 exp.DayOfWeek: rename_func("DAYOFWEEK"), 177 exp.DayOfYear: rename_func("DAYOFYEAR"), 178 exp.DataType: _datatype_sql, 179 exp.DateAdd: _date_add_sql, 180 exp.DateDiff: lambda self, e: self.func( 181 "DATE_DIFF", f"'{e.args.get('unit', 'day')}'", e.expression, e.this 182 ), 183 exp.DateStrToDate: datestrtodate_sql, 184 exp.DateToDi: lambda self, e: f"CAST(STRFTIME({self.sql(e, 'this')}, {DuckDB.dateint_format}) AS INT)", 185 exp.DiToDate: lambda self, e: f"CAST(STRPTIME(CAST({self.sql(e, 'this')} AS TEXT), {DuckDB.dateint_format}) AS DATE)", 186 exp.Explode: rename_func("UNNEST"), 187 exp.JSONExtract: arrow_json_extract_sql, 188 exp.JSONExtractScalar: arrow_json_extract_scalar_sql, 189 exp.JSONBExtract: arrow_json_extract_sql, 190 exp.JSONBExtractScalar: arrow_json_extract_scalar_sql, 191 exp.LogicalOr: rename_func("BOOL_OR"), 192 exp.LogicalAnd: rename_func("BOOL_AND"), 193 exp.Pivot: no_pivot_sql, 194 exp.Properties: no_properties_sql, 195 exp.RegexpExtract: _regexp_extract_sql, 196 exp.RegexpLike: rename_func("REGEXP_MATCHES"), 197 exp.RegexpSplit: rename_func("STR_SPLIT_REGEX"), 198 exp.SafeDivide: no_safe_divide_sql, 199 exp.Split: rename_func("STR_SPLIT"), 200 exp.SortArray: _sort_array_sql, 201 exp.StrPosition: str_position_sql, 202 exp.StrToDate: lambda self, e: f"CAST({str_to_time_sql(self, e)} AS DATE)", 203 exp.StrToTime: str_to_time_sql, 204 exp.StrToUnix: lambda self, e: f"EPOCH(STRPTIME({self.sql(e, 'this')}, {self.format_time(e)}))", 205 exp.Struct: _struct_sql, 206 exp.TimestampTrunc: timestamptrunc_sql, 207 exp.TimeStrToDate: lambda self, e: f"CAST({self.sql(e, 'this')} AS DATE)", 208 exp.TimeStrToTime: timestrtotime_sql, 209 exp.TimeStrToUnix: lambda self, e: f"EPOCH(CAST({self.sql(e, 'this')} AS TIMESTAMP))", 210 exp.TimeToStr: lambda self, e: f"STRFTIME({self.sql(e, 'this')}, {self.format_time(e)})", 211 exp.TimeToUnix: rename_func("EPOCH"), 212 exp.TsOrDiToDi: lambda self, e: f"CAST(SUBSTR(REPLACE(CAST({self.sql(e, 'this')} AS TEXT), '-', ''), 1, 8) AS INT)", 213 exp.TsOrDsAdd: _ts_or_ds_add_sql, 214 exp.TsOrDsToDate: ts_or_ds_to_date_sql("duckdb"), 215 exp.UnixToStr: lambda self, e: f"STRFTIME(TO_TIMESTAMP({self.sql(e, 'this')}), {self.format_time(e)})", 216 exp.UnixToTime: rename_func("TO_TIMESTAMP"), 217 exp.UnixToTimeStr: lambda self, e: f"CAST(TO_TIMESTAMP({self.sql(e, 'this')}) AS TEXT)", 218 exp.WeekOfYear: rename_func("WEEKOFYEAR"), 219 } 220 221 TYPE_MAPPING = { 222 **generator.Generator.TYPE_MAPPING, 223 exp.DataType.Type.BINARY: "BLOB", 224 exp.DataType.Type.CHAR: "TEXT", 225 exp.DataType.Type.FLOAT: "REAL", 226 exp.DataType.Type.NCHAR: "TEXT", 227 exp.DataType.Type.NVARCHAR: "TEXT", 228 exp.DataType.Type.UINT: "UINTEGER", 229 exp.DataType.Type.VARBINARY: "BLOB", 230 exp.DataType.Type.VARCHAR: "TEXT", 231 } 232 233 STAR_MAPPING = {**generator.Generator.STAR_MAPPING, "except": "EXCLUDE"} 234 235 PROPERTIES_LOCATION = { 236 **generator.Generator.PROPERTIES_LOCATION, # type: ignore 237 exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED, 238 } 239 240 def tablesample_sql( 241 self, expression: exp.TableSample, seed_prefix: str = "SEED", sep=" AS " 242 ) -> str: 243 return super().tablesample_sql(expression, seed_prefix="REPEATABLE", sep=sep)
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', sep=' AS ') -> str:
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
- windowspec_sql
- withingroup_sql
- between_sql
- bracket_sql
- all_sql
- any_sql
- exists_sql
- case_sql
- constraint_sql
- nextvaluefor_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