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