SNOW-2155042 Add --format JSON_EXT that will expand nested JSONs (#2458)

* SNOW-2155042 Properly display nested JSONs

* update RN

* fix integration test

* fix mocks

* Use ENUMs

* revert test changes

* make json expanding opt-in

* make expand_json globally available

* ambr files, but I'm not sure if we shouldn't add format JSON_2 or something

* one more

* Replace --expand-json flag with JSON_EXT format option

* ambr

* review fixes

Author: sfc-gh-daniszewski

RELEASE-NOTES.md

@@ -43,6 +43,7 @@
 ## Fixes and improvements
 * Fixed failing snow sql command when snowflake.yml is invalid and query has no templating.
 * Fix JSON serialization for `Decimal`, `time` and `binary`.
+* Added `--format=JSON_EXT` option to return JSON objects as proper JSON structures rather than strings.
 * Refactored Streamlit app deployment (using `FROM <stage>` syntax); removed deprecated Streamlit features
 
 

src/snowflake/cli/_app/printing.py

@@ -104,7 +104,7 @@ def __to_str(val):
 
 
 def is_structured_format(output_format):
-    return output_format in [OutputFormat.JSON, OutputFormat.CSV]
+    return output_format.is_json or output_format == OutputFormat.CSV
 
 
 def print_structured(

src/snowflake/cli/_plugins/nativeapp/commands.py

@@ -55,7 +55,6 @@
     IncompatibleParametersError,
     UnmetParametersError,
 )
-from snowflake.cli.api.output.formats import OutputFormat
 from snowflake.cli.api.output.types import (
     CommandResult,
     MessageResult,
@@ -117,9 +116,9 @@ def app_diff(
     diff = ws.perform_action(
         package_id,
         EntityActions.DIFF,
-        print_to_console=cli_context.output_format != OutputFormat.JSON,
+        print_to_console=not cli_context.output_format.is_json,
     )
-    if cli_context.output_format == OutputFormat.JSON:
+    if cli_context.output_format.is_json:
         return ObjectResult(diff.to_dict())
 
     return None
@@ -373,7 +372,7 @@ def app_validate(
     )
     package_id = options["package_entity_id"]
     package = ws.get_entity(package_id)
-    if cli_context.output_format == OutputFormat.JSON:
+    if cli_context.output_format.is_json:
         return ObjectResult(
             package.get_validation_result(
                 action_ctx=ws.action_ctx,

src/snowflake/cli/_plugins/nativeapp/release_channel/commands.py

@@ -26,7 +26,6 @@
 from snowflake.cli.api.commands.decorators import with_project_definition
 from snowflake.cli.api.commands.snow_typer import SnowTyperFactory
 from snowflake.cli.api.entities.utils import EntityActions
-from snowflake.cli.api.output.formats import OutputFormat
 from snowflake.cli.api.output.types import (
     CollectionResult,
     CommandResult,
@@ -68,7 +67,7 @@ def release_channel_list(
         release_channel=channel,
     )
 
-    if cli_context.output_format == OutputFormat.JSON:
+    if cli_context.output_format.is_json:
         return CollectionResult(channels)
 
 

src/snowflake/cli/_plugins/nativeapp/version/commands.py

@@ -30,7 +30,6 @@
 )
 from snowflake.cli.api.commands.snow_typer import SnowTyperFactory
 from snowflake.cli.api.entities.utils import EntityActions
-from snowflake.cli.api.output.formats import OutputFormat
 from snowflake.cli.api.output.types import (
     CollectionResult,
     CommandResult,
@@ -105,7 +104,7 @@ def create(
     )
 
     message = "Version create is now complete."
-    if cli_context.output_format == OutputFormat.JSON:
+    if cli_context.output_format.is_json:
         return ObjectResult(
             {
                 "message": message,

src/snowflake/cli/_plugins/stage/commands.py

@@ -48,7 +48,6 @@
 from snowflake.cli.api.console import cli_console
 from snowflake.cli.api.constants import ObjectType
 from snowflake.cli.api.identifiers import FQN
-from snowflake.cli.api.output.formats import OutputFormat
 from snowflake.cli.api.output.types import (
     CollectionResult,
     CommandResult,
@@ -219,7 +218,7 @@ def stage_diff(
         local_root=Path(folder_name),
         stage_path=StageManager.stage_path_parts_from_str(stage_name),  # noqa: SLF001
     )
-    if get_cli_context().output_format == OutputFormat.JSON:
+    if get_cli_context().output_format.is_json:
         return ObjectResult(diff.to_dict())
     else:
         print_diff_to_console(diff)

src/snowflake/cli/api/cli_global_context.py

@@ -196,7 +196,10 @@ def silent(self) -> bool:
     @property
     def _should_force_mute_intermediate_output(self) -> bool:
         """Computes whether cli_console output should be muted."""
-        return self._manager.output_format in [OutputFormat.JSON, OutputFormat.CSV]
+        return (
+            self._manager.output_format.is_json
+            or self._manager.output_format == OutputFormat.CSV
+        )
 
     @property
     def enhanced_exit_codes(self) -> bool:

src/snowflake/cli/api/commands/flags.py

@@ -446,6 +446,7 @@ def _diag_log_allowlist_path_callback(path: str):
     rich_help_panel=_CLI_BEHAVIOUR,
 )
 
+
 SilentOption = typer.Option(
     False,
     "--silent",

src/snowflake/cli/api/output/formats.py

@@ -18,4 +18,9 @@
 class OutputFormat(Enum):
     TABLE = "TABLE"
     JSON = "JSON"
+    JSON_EXT = "JSON_EXT"
     CSV = "CSV"
+
+    @property
+    def is_json(self) -> bool:
+        return self in (OutputFormat.JSON, OutputFormat.JSON_EXT)

src/snowflake/cli/api/output/types.py

@@ -16,11 +16,22 @@
 
 import json
 import typing as t
+from enum import IntEnum
 
+from snowflake.cli.api.cli_global_context import get_cli_context
+from snowflake.cli.api.output.formats import OutputFormat
 from snowflake.connector import DictCursor
 from snowflake.connector.cursor import SnowflakeCursor
 
 
+class SnowflakeColumnType(IntEnum):
+    """Snowflake column type codes for JSON-capable data types."""
+
+    VARIANT = 5
+    OBJECT = 9
+    ARRAY = 10
+
+
 class CommandResult:
     @property
     def result(self):
@@ -69,13 +80,48 @@ def result(self):
 class QueryResult(CollectionResult):
     def __init__(self, cursor: SnowflakeCursor | DictCursor):
         self.column_names = [col.name for col in cursor.description]
+        # Store column type information to identify VARIANT columns (JSON data)
+        self.column_types = [col.type_code for col in cursor.description]
         super().__init__(elements=self._prepare_payload(cursor))
         self._query = cursor.query
 
     def _prepare_payload(self, cursor: SnowflakeCursor | DictCursor):
         if isinstance(cursor, DictCursor):
-            return (k for k in cursor)
-        return ({k: v for k, v in zip(self.column_names, row)} for row in cursor)
+            return (self._process_columns(k) for k in cursor)
+        return (
+            self._process_columns({k: v for k, v in zip(self.column_names, row)})
+            for row in cursor
+        )
+
+    def _process_columns(self, row_dict):
+        if get_cli_context().output_format != OutputFormat.JSON_EXT:
+            return row_dict
+
+        processed_row = {}
+        for i, (column_name, value) in enumerate(row_dict.items()):
+            # Check if this column can contain JSON data
+            if i < len(self.column_types) and self.column_types[i] in (
+                SnowflakeColumnType.VARIANT,
+                SnowflakeColumnType.OBJECT,
+                SnowflakeColumnType.ARRAY,
+            ):
+                # For ARRAY and OBJECT types, the values are always JSON strings that need parsing
+                # For VARIANT types, only parse if the value is a string
+                if self.column_types[i] in (
+                    SnowflakeColumnType.OBJECT,
+                    SnowflakeColumnType.ARRAY,
+                ) or isinstance(value, str):
+                    try:
+                        # Try to parse as JSON
+                        processed_row[column_name] = json.loads(value)
+                    except (json.JSONDecodeError, TypeError):
+                        # If parsing fails, keep the original value
+                        processed_row[column_name] = value
+                else:
+                    processed_row[column_name] = value
+            else:
+                processed_row[column_name] = value
+        return processed_row
 
     @property
     def query(self):