Merging upstream version 16.4.0.

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

1 files changed, 22 insertions, 0 deletions

diff --git a/sqlglot/dialects/bigquery.py b/sqlglot/dialects/bigquery.py
index 2166e65..52d4a88 100644
--- a/sqlglot/dialects/bigquery.py
+++ b/sqlglot/dialects/bigquery.py
@@ -4,6 +4,7 @@ import re
 import typing as t
 
 from sqlglot import exp, generator, parser, tokens, transforms
+from sqlglot._typing import E
 from sqlglot.dialects.dialect import (
     Dialect,
     datestrtodate_sql,
@@ -106,6 +107,9 @@ def _unqualify_unnest(expression: exp.Expression) -> exp.Expression:
 class BigQuery(Dialect):
     UNNEST_COLUMN_ONLY = True
 
+    # https://cloud.google.com/bigquery/docs/reference/standard-sql/lexical#case_sensitivity
+    RESOLVES_IDENTIFIERS_AS_UPPERCASE = None
+
     TIME_MAPPING = {
         "%D": "%m/%d/%y",
     }
@@ -126,6 +130,20 @@ class BigQuery(Dialect):
         "TZH": "%z",
     }
 
+    @classmethod
+    def normalize_identifier(cls, expression: E) -> E:
+        # In BigQuery, CTEs aren't case-sensitive, but table names are (by default, at least).
+        # The following check is essentially a heuristic to detect tables based on whether or
+        # not they're qualified.
+        if (
+            isinstance(expression, exp.Identifier)
+            and not (isinstance(expression.parent, exp.Table) and expression.parent.db)
+            and not expression.meta.get("is_table")
+        ):
+            expression.set("this", expression.this.lower())
+
+        return expression
+
     class Tokenizer(tokens.Tokenizer):
         QUOTES = ["'", '"', '"""', "'''"]
         COMMENTS = ["--", "#", ("/*", "*/")]
@@ -176,6 +194,7 @@ class BigQuery(Dialect):
             "DATETIME_ADD": parse_date_delta_with_interval(exp.DatetimeAdd),
             "DATETIME_SUB": parse_date_delta_with_interval(exp.DatetimeSub),
             "DIV": lambda args: exp.IntDiv(this=seq_get(args, 0), expression=seq_get(args, 1)),
+            "GENERATE_ARRAY": exp.GenerateSeries.from_arg_list,
             "PARSE_DATE": lambda args: format_time_lambda(exp.StrToDate, "bigquery")(
                 [seq_get(args, 1), seq_get(args, 0)]
             ),
@@ -201,6 +220,7 @@ class BigQuery(Dialect):
             "TIME_SUB": parse_date_delta_with_interval(exp.TimeSub),
             "TIMESTAMP_ADD": parse_date_delta_with_interval(exp.TimestampAdd),
             "TIMESTAMP_SUB": parse_date_delta_with_interval(exp.TimestampSub),
+            "TO_JSON_STRING": exp.JSONFormat.from_arg_list,
         }
 
         FUNCTION_PARSERS = {
@@ -289,6 +309,8 @@ class BigQuery(Dialect):
             exp.DateDiff: lambda self, e: f"DATE_DIFF({self.sql(e, 'this')}, {self.sql(e, 'expression')}, {self.sql(e.args.get('unit', 'DAY'))})",
             exp.DateStrToDate: datestrtodate_sql,
             exp.DateTrunc: lambda self, e: self.func("DATE_TRUNC", e.this, e.text("unit")),
+            exp.JSONFormat: rename_func("TO_JSON_STRING"),
+            exp.GenerateSeries: rename_func("GENERATE_ARRAY"),
             exp.GroupConcat: rename_func("STRING_AGG"),
             exp.ILike: no_ilike_sql,
             exp.IntDiv: rename_func("DIV"),