Previous value: -"Machine-readable failure mode. Declared by this tool: `canvas_unavailable`: DataCanvas is not configured in this deployment. `missing_table`: A table the SQL names is not staged — a df_ name that expired, was dropped to make room for newer dataframes, or is mistyped, or any other table name. `invalid_sql`: A SELECT fails to parse or prepare (syntax error, unknown column or function, bad expression). `sql_execution_error`: SELECT prepared but failed on the data. `non_select_statement`: The statement is not a SELECT, or cannot be parsed as one. `multi_statement`: The SQL holds more than one statement. `denied_function`: The SQL calls a file-reading, external-data, engine-setting, or dynamic SQL function the read-only gate refuses. `plan_operator_not_allowed`: The query plan uses an operator outside the read-only allowlist (scans of staged tables, filters, joins, aggregates, set operations, sorts, window functions, CTEs, unnest). `system_catalog_access`: The SQL references a system catalog: information_schema, pg_catalog, sqlite_master, or a duckdb_*() function. `register_as_clash`: register_as names a dataframe that is already staged. `register_as_too_large`: The register_as result holds more rows than the staging budget (1,000,000 rows) allows on its own. Other values are possible when a failure originates below the handler."New value: +"Machine-readable failure mode. Declared by this tool: `canvas_unavailable`: DataCanvas is not configured in this deployment. `missing_table`: A table the SQL names is not staged — a df_ name that expired, was dropped to make room for newer dataframes, or is mistyped, or any other table name. `invalid_sql`: A SELECT fails to parse or prepare (syntax error, unknown column or function, bad expression). `sql_execution_error`: SELECT prepared but failed on the data. `non_select_statement`: The statement is not a SELECT, or cannot be parsed as one. `multi_statement`: The SQL holds more than one statement. `denied_function`: The SQL calls a file-reading, external-data, engine-setting, or dynamic SQL function the read-only gate refuses, or names current_schema unqualified, which DuckDB reads as the current_schema() function. `plan_operator_not_allowed`: The query plan uses an operator outside the read-only allowlist (scans of staged tables, filters, joins, aggregates, set operations, sorts, window functions, CTEs, unnest). `system_catalog_access`: The SQL references a system catalog: information_schema, pg_catalog, sqlite_master, or a duckdb_*() function. `register_as_clash`: register_as names a dataframe that is already staged. `register_as_too_large`: The register_as result holds more rows than the staging budget (1,000,000 rows) allows on its own. Other values are possible when a failure originates below the handler."

Previous value: -"Rows to return inline (0–10,000); defaults to row_limit, and a value above row_limit is treated as row_limit. Set it low when register_as keeps the full result."New value: +"Rows to return inline; defaults to the effective row_limit and is clamped to it. Set it low when register_as keeps the full result."

Previous value: -10000New value: +9007199254740991

Previous value: -"Hard cap on rows materialized (1–10,000). A query matching more stops here and row_count_capped comes back true; use register_as to keep the whole result."New value: +"Requested materialization cap (default 1,000; hard input ceiling 1,000,000), clamped to the execution cap of 10,000. Without register_as, row_count_capped reports more matches; continue with ordered LIMIT/OFFSET pages. register_as keeps the whole result within the staging budget."

Previous value: -10000New value: +1000000

Previous value: -[
-  {
-    "not": {
-      "required": [
-        "error"
-      ]
-    },
-    "required": [
-      "columns",
-      "row_count",
-      "row_count_capped",
-      "rows"
-    ]
-  },
-  {
-    "required": [
-      "error"
-    ]
-  }
-]New value: +[
+  {
+    "not": {
+      "required": [
+        "error"
+      ]
+    },
+    "required": [
+      "requested_row_limit",
+      "effective_row_limit",
+      "effective_preview",
+      "columns",
+      "row_count",
+      "row_count_capped",
+      "rows"
+    ]
+  },
+  {
+    "required": [
+      "error"
+    ]
+  }
+]

Added value: +{
+  "description": "Inline row cap after clamping preview to the effective row_limit.",
+  "maximum": 9007199254740991,
+  "minimum": -9007199254740991,
+  "type": "integer"
+}

Added value: +{
+  "description": "Effective row_limit, at most 10,000.",
+  "maximum": 9007199254740991,
+  "minimum": -9007199254740991,
+  "type": "integer"
+}

Previous value: -"Machine-readable failure mode. Declared by this tool: `canvas_unavailable`: DataCanvas is not configured in this deployment. `missing_table`: A table the SQL names is not staged — a df_ name that expired, was dropped to make room for newer dataframes, or is mistyped, or any other table name. `invalid_sql`: A SELECT fails to parse or prepare (syntax error, unknown column or function, bad expression). `sql_execution_error`: SELECT prepared but failed on the data. `non_select_statement`: The statement is not a SELECT, or cannot be parsed as one. `multi_statement`: The SQL holds more than one statement. `denied_function`: The SQL calls a file-reading or external-data table function such as read_csv or read_parquet. `plan_operator_not_allowed`: The query plan uses an operator outside the read-only allowlist (scans of staged tables, filters, joins, aggregates, set operations, sorts, window functions, CTEs, unnest). `system_catalog_access`: The SQL references a system catalog: information_schema, pg_catalog, sqlite_master, or a duckdb_*() function. `register_as_clash`: register_as names a dataframe that is already staged. `register_as_too_large`: The register_as result holds more rows than the staging budget (1,000,000 rows) allows on its own. Other values are possible when a failure originates below the handler."New value: +"Machine-readable failure mode. Declared by this tool: `canvas_unavailable`: DataCanvas is not configured in this deployment. `missing_table`: A table the SQL names is not staged — a df_ name that expired, was dropped to make room for newer dataframes, or is mistyped, or any other table name. `invalid_sql`: A SELECT fails to parse or prepare (syntax error, unknown column or function, bad expression). `sql_execution_error`: SELECT prepared but failed on the data. `non_select_statement`: The statement is not a SELECT, or cannot be parsed as one. `multi_statement`: The SQL holds more than one statement. `denied_function`: The SQL calls a file-reading, external-data, engine-setting, or dynamic SQL function the read-only gate refuses. `plan_operator_not_allowed`: The query plan uses an operator outside the read-only allowlist (scans of staged tables, filters, joins, aggregates, set operations, sorts, window functions, CTEs, unnest). `system_catalog_access`: The SQL references a system catalog: information_schema, pg_catalog, sqlite_master, or a duckdb_*() function. `register_as_clash`: register_as names a dataframe that is already staged. `register_as_too_large`: The register_as result holds more rows than the staging budget (1,000,000 rows) allows on its own. Other values are possible when a failure originates below the handler."

Added value: +{
+  "description": "Explicit preview before clamping; absent when omitted.",
+  "maximum": 9007199254740991,
+  "minimum": -9007199254740991,
+  "type": "integer"
+}

Added value: +{
+  "description": "Requested row_limit before the execution cap.",
+  "maximum": 9007199254740991,
+  "minimum": -9007199254740991,
+  "type": "integer"
+}

Previous value: -"Rows the query produced, up to row_limit; with register_as, the exact count of the new dataframe."New value: +"Rows the query produced, up to effective_row_limit; with register_as, the exact count of the new dataframe."

Previous value: -"True when the query matched more rows than row_limit, so row_count is that cap."New value: +"True when the query matched more rows than effective_row_limit, so row_count is that cap."