diff options
Diffstat (limited to 'sqlglot/dialects/hive.py')
| -rw-r--r-- | sqlglot/dialects/hive.py | 23 |
1 files changed, 15 insertions, 8 deletions
diff --git a/sqlglot/dialects/hive.py b/sqlglot/dialects/hive.py index 8847119..6bca610 100644 --- a/sqlglot/dialects/hive.py +++ b/sqlglot/dialects/hive.py @@ -86,13 +86,17 @@ def _date_diff_sql(self: generator.Generator, expression: exp.DateDiff) -> str: def _json_format_sql(self: generator.Generator, expression: exp.JSONFormat) -> str: this = expression.this - if not this.type: - from sqlglot.optimizer.annotate_types import annotate_types + if isinstance(this, exp.Cast) and this.is_type("json") and this.this.is_string: + # Since FROM_JSON requires a nested type, we always wrap the json string with + # an array to ensure that "naked" strings like "'a'" will be handled correctly + wrapped_json = exp.Literal.string(f"[{this.this.name}]") - annotate_types(this) + from_json = self.func("FROM_JSON", wrapped_json, self.func("SCHEMA_OF_JSON", wrapped_json)) + to_json = self.func("TO_JSON", from_json) + + # This strips the [, ] delimiters of the dummy array printed by TO_JSON + return self.func("REGEXP_EXTRACT", to_json, "'^.(.*).$'", "1") - if this.type.is_type("json"): - return self.sql(this) return self.func("TO_JSON", this, expression.args.get("options")) @@ -153,6 +157,9 @@ class Hive(Dialect): ALIAS_POST_TABLESAMPLE = True IDENTIFIERS_CAN_START_WITH_DIGIT = True + # https://spark.apache.org/docs/latest/sql-ref-identifier.html#description + RESOLVES_IDENTIFIERS_AS_UPPERCASE = None + TIME_MAPPING = { "y": "%Y", "Y": "%Y", @@ -268,9 +275,9 @@ class Hive(Dialect): QUERY_MODIFIER_PARSERS = { **parser.Parser.QUERY_MODIFIER_PARSERS, - "distribute": lambda self: self._parse_sort(exp.Distribute, "DISTRIBUTE", "BY"), - "sort": lambda self: self._parse_sort(exp.Sort, "SORT", "BY"), - "cluster": lambda self: self._parse_sort(exp.Cluster, "CLUSTER", "BY"), + "cluster": lambda self: self._parse_sort(exp.Cluster, TokenType.CLUSTER_BY), + "distribute": lambda self: self._parse_sort(exp.Distribute, TokenType.DISTRIBUTE_BY), + "sort": lambda self: self._parse_sort(exp.Sort, TokenType.SORT_BY), } def _parse_types( |