sqlglot.dialects.snowflake

 17class Snowflake(Dialect):
 18    # https://docs.snowflake.com/en/sql-reference/identifiers-syntax
 19    NORMALIZATION_STRATEGY = NormalizationStrategy.UPPERCASE
 20    NULL_ORDERING = "nulls_are_large"
 21    TIME_FORMAT = "'YYYY-MM-DD HH24:MI:SS'"
 22    SUPPORTS_USER_DEFINED_TYPES = False
 23    PREFER_CTE_ALIAS_COLUMN = True
 24    TABLESAMPLE_SIZE_IS_PERCENT = True
 25    COPY_PARAMS_ARE_CSV = False
 26    ARRAY_AGG_INCLUDES_NULLS = None
 27    ARRAY_FUNCS_PROPAGATES_NULLS = True
 28    ALTER_TABLE_ADD_REQUIRED_FOR_EACH_COLUMN = False
 29    TRY_CAST_REQUIRES_STRING = True
 30    SUPPORTS_ALIAS_REFS_IN_JOIN_CONDITIONS = True
 31    LEAST_GREATEST_IGNORES_NULLS = False
 32
 33    EXPRESSION_METADATA = EXPRESSION_METADATA.copy()
 34
 35    # https://docs.snowflake.com/en/en/sql-reference/functions/initcap
 36    INITCAP_DEFAULT_DELIMITER_CHARS = ' \t\n\r\f\v!?@"^#$&~_,.:;+\\-*%/|\\[\\](){}<>'
 37
 38    INVERSE_TIME_MAPPING = {
 39        "T": "T",  # in TIME_MAPPING we map '"T"' with the double quotes to 'T', and we want to prevent 'T' from being mapped back to '"T"' so that 'AUTO' doesn't become 'AU"T"O'
 40    }
 41
 42    TIME_MAPPING = {
 43        "YYYY": "%Y",
 44        "yyyy": "%Y",
 45        "YY": "%y",
 46        "yy": "%y",
 47        "MMMM": "%B",
 48        "mmmm": "%B",
 49        "MON": "%b",
 50        "mon": "%b",
 51        "MM": "%m",
 52        "mm": "%m",
 53        "DD": "%d",
 54        "dd": "%-d",
 55        "DY": "%a",
 56        "dy": "%w",
 57        "HH24": "%H",
 58        "hh24": "%H",
 59        "HH12": "%I",
 60        "hh12": "%I",
 61        "MI": "%M",
 62        "mi": "%M",
 63        "SS": "%S",
 64        "ss": "%S",
 65        "FF": "%f_nine",  # %f_ internal representation with precision specified
 66        "ff": "%f_nine",
 67        "FF0": "%f_zero",
 68        "ff0": "%f_zero",
 69        "FF1": "%f_one",
 70        "ff1": "%f_one",
 71        "FF2": "%f_two",
 72        "ff2": "%f_two",
 73        "FF3": "%f_three",
 74        "ff3": "%f_three",
 75        "FF4": "%f_four",
 76        "ff4": "%f_four",
 77        "FF5": "%f_five",
 78        "ff5": "%f_five",
 79        "FF6": "%f",
 80        "ff6": "%f",
 81        "FF7": "%f_seven",
 82        "ff7": "%f_seven",
 83        "FF8": "%f_eight",
 84        "ff8": "%f_eight",
 85        "FF9": "%f_nine",
 86        "ff9": "%f_nine",
 87        "TZHTZM": "%z",
 88        "tzhtzm": "%z",
 89        "TZH:TZM": "%:z",  # internal representation for ±HH:MM
 90        "tzh:tzm": "%:z",
 91        "TZH": "%-z",  # internal representation ±HH
 92        "tzh": "%-z",
 93        '"T"': "T",  # remove the optional double quotes around the separator between the date and time
 94        # Seems like Snowflake treats AM/PM in the format string as equivalent,
 95        # only the time (stamp) value's AM/PM affects the output
 96        "AM": "%p",
 97        "am": "%p",
 98        "PM": "%p",
 99        "pm": "%p",
100    }
101
102    DATE_PART_MAPPING = {
103        **Dialect.DATE_PART_MAPPING,
104        "ISOWEEK": "WEEKISO",
105        # The base Dialect maps EPOCH_SECOND -> EPOCH, but we need to preserve
106        # EPOCH_SECOND as a distinct value for two reasons:
107        # 1. Type annotation: EPOCH_SECOND returns BIGINT, while EPOCH returns DOUBLE
108        # 2. Transpilation: DuckDB's EPOCH() returns float, so we cast EPOCH_SECOND
109        #    to BIGINT to match Snowflake's integer behavior
110        # Without this override, EXTRACT(EPOCH_SECOND FROM ts) would be normalized
111        # to EXTRACT(EPOCH FROM ts) and lose the integer semantics.
112        "EPOCH_SECOND": "EPOCH_SECOND",
113        "EPOCH_SECONDS": "EPOCH_SECOND",
114    }
115
116    PSEUDOCOLUMNS = {"LEVEL"}
117
118    def can_quote(self, identifier: exp.Identifier, identify: str | bool = "safe") -> bool:
119        # This disables quoting DUAL in SELECT ... FROM DUAL, because Snowflake treats an
120        # unquoted DUAL keyword in a special way and does not map it to a user-defined table
121        return super().can_quote(identifier, identify) and not (
122            isinstance(identifier.parent, exp.Table)
123            and not identifier.quoted
124            and identifier.name.lower() == "dual"
125        )
126
127    class JSONPathTokenizer(jsonpath.JSONPathTokenizer):
128        SINGLE_TOKENS = jsonpath.JSONPathTokenizer.SINGLE_TOKENS.copy()
129        SINGLE_TOKENS.pop("$")
130
131    Parser = SnowflakeParser
132
133    class Tokenizer(tokens.Tokenizer):
134        STRING_ESCAPES = ["\\", "'"]
135        HEX_STRINGS = [("x'", "'"), ("X'", "'")]
136        RAW_STRINGS = ["$$"]
137        COMMENTS = ["--", "//", ("/*", "*/")]
138        NESTED_COMMENTS = False
139
140        KEYWORDS = {
141            **tokens.Tokenizer.KEYWORDS,
142            "BYTEINT": TokenType.INT,
143            "FILE://": TokenType.URI_START,
144            "FILE FORMAT": TokenType.FILE_FORMAT,
145            "GET": TokenType.GET,
146            "INTEGRATION": TokenType.INTEGRATION,
147            "MATCH_CONDITION": TokenType.MATCH_CONDITION,
148            "MATCH_RECOGNIZE": TokenType.MATCH_RECOGNIZE,
149            "MINUS": TokenType.EXCEPT,
150            "NCHAR VARYING": TokenType.VARCHAR,
151            "PACKAGE": TokenType.PACKAGE,
152            "POLICY": TokenType.POLICY,
153            "POOL": TokenType.POOL,
154            "PUT": TokenType.PUT,
155            "REMOVE": TokenType.COMMAND,
156            "RM": TokenType.COMMAND,
157            "ROLE": TokenType.ROLE,
158            "RULE": TokenType.RULE,
159            "SAMPLE": TokenType.TABLE_SAMPLE,
160            "SEMANTIC VIEW": TokenType.SEMANTIC_VIEW,
161            "SQL_DOUBLE": TokenType.DOUBLE,
162            "SQL_VARCHAR": TokenType.VARCHAR,
163            "STAGE": TokenType.STAGE,
164            "STORAGE INTEGRATION": TokenType.STORAGE_INTEGRATION,
165            "STREAMLIT": TokenType.STREAMLIT,
166            "TAG": TokenType.TAG,
167            "TIMESTAMP_TZ": TokenType.TIMESTAMPTZ,
168            "TOP": TokenType.TOP,
169            "VOLUME": TokenType.VOLUME,
170            "WAREHOUSE": TokenType.WAREHOUSE,
171            # https://docs.snowflake.com/en/sql-reference/data-types-numeric#float
172            # FLOAT is a synonym for DOUBLE in Snowflake
173            "FLOAT": TokenType.DOUBLE,
174        }
175        KEYWORDS.pop("/*+")
176
177        SINGLE_TOKENS = {
178            **tokens.Tokenizer.SINGLE_TOKENS,
179            "$": TokenType.PARAMETER,
180            "!": TokenType.EXCLAMATION,
181        }
182
183        VAR_SINGLE_TOKENS = {"$"}
184
185        COMMANDS = tokens.Tokenizer.COMMANDS - {TokenType.SHOW}
186
187    Generator = SnowflakeGenerator

127    class JSONPathTokenizer(jsonpath.JSONPathTokenizer):
128        SINGLE_TOKENS = jsonpath.JSONPathTokenizer.SINGLE_TOKENS.copy()
129        SINGLE_TOKENS.pop("$")

133    class Tokenizer(tokens.Tokenizer):
134        STRING_ESCAPES = ["\\", "'"]
135        HEX_STRINGS = [("x'", "'"), ("X'", "'")]
136        RAW_STRINGS = ["$$"]
137        COMMENTS = ["--", "//", ("/*", "*/")]
138        NESTED_COMMENTS = False
139
140        KEYWORDS = {
141            **tokens.Tokenizer.KEYWORDS,
142            "BYTEINT": TokenType.INT,
143            "FILE://": TokenType.URI_START,
144            "FILE FORMAT": TokenType.FILE_FORMAT,
145            "GET": TokenType.GET,
146            "INTEGRATION": TokenType.INTEGRATION,
147            "MATCH_CONDITION": TokenType.MATCH_CONDITION,
148            "MATCH_RECOGNIZE": TokenType.MATCH_RECOGNIZE,
149            "MINUS": TokenType.EXCEPT,
150            "NCHAR VARYING": TokenType.VARCHAR,
151            "PACKAGE": TokenType.PACKAGE,
152            "POLICY": TokenType.POLICY,
153            "POOL": TokenType.POOL,
154            "PUT": TokenType.PUT,
155            "REMOVE": TokenType.COMMAND,
156            "RM": TokenType.COMMAND,
157            "ROLE": TokenType.ROLE,
158            "RULE": TokenType.RULE,
159            "SAMPLE": TokenType.TABLE_SAMPLE,
160            "SEMANTIC VIEW": TokenType.SEMANTIC_VIEW,
161            "SQL_DOUBLE": TokenType.DOUBLE,
162            "SQL_VARCHAR": TokenType.VARCHAR,
163            "STAGE": TokenType.STAGE,
164            "STORAGE INTEGRATION": TokenType.STORAGE_INTEGRATION,
165            "STREAMLIT": TokenType.STREAMLIT,
166            "TAG": TokenType.TAG,
167            "TIMESTAMP_TZ": TokenType.TIMESTAMPTZ,
168            "TOP": TokenType.TOP,
169            "VOLUME": TokenType.VOLUME,
170            "WAREHOUSE": TokenType.WAREHOUSE,
171            # https://docs.snowflake.com/en/sql-reference/data-types-numeric#float
172            # FLOAT is a synonym for DOUBLE in Snowflake
173            "FLOAT": TokenType.DOUBLE,
174        }
175        KEYWORDS.pop("/*+")
176
177        SINGLE_TOKENS = {
178            **tokens.Tokenizer.SINGLE_TOKENS,
179            "$": TokenType.PARAMETER,
180            "!": TokenType.EXCLAMATION,
181        }
182
183        VAR_SINGLE_TOKENS = {"$"}
184
185        COMMANDS = tokens.Tokenizer.COMMANDS - {TokenType.SHOW}