Hyper: Regenerate queries with new operator / expression names

plan-dumper/README.md

@@ -27,6 +27,7 @@ The available modes are:
 * `simple` → a plain JSON `EXPLAIN`.
 * `analyze` → an executing plan with runtime measurements.
   Hyper, Postgres, DuckDB, Umbra, CedarDB, and MariaDB use their JSON `EXPLAIN ANALYZE` variants.
+* `external-analyze` → Hyper's external JSON executing plan (`FORMAT JSON, ANALYZE`).
 * `steps` → optimizer internals:
   * Hyper: `EXPLAIN (FORMAT INTERNAL, OPTIMIZE STEPS)`.
   * DuckDB: logical, optimized, and physical plans from `explain_output='all'`.

plan-dumper/dump-plans.py

@@ -44,7 +44,7 @@ def parse_unsupported(sql):
     return {db.strip().lower() for db in match.group(1).split(",")}
 
 
-# A `-- MODES: simple, analyze, pipelines` comment anywhere in a query file lists the
+# A `-- MODES: simple, analyze, external-analyze, pipelines` comment anywhere in a query file lists the
 # EXPLAIN modes to dump it under (one output file per mode); defaults to `analyze` alone
 # when absent, since that's what most queries want.
 modes_re = re.compile(r"^--\s*MODES:\s*(.+)$", re.MULTILINE | re.IGNORECASE)
@@ -341,6 +341,7 @@ def get_plan(sql, mode):
                     "simple": "FORMAT INTERNAL",
                     "steps": "FORMAT INTERNAL, OPTIMIZE STEPS",
                     "analyze": "FORMAT INTERNAL, ANALYZE",
+                    "external-analyze": "FORMAT JSON, ANALYZE, EXPAND_VIEWS true",
                     "pipelines": (
                         "FORMAT INTERNAL, PIPELINES, EXPAND_VIEWS true, EXPRESSIONS SQL"
                     ),

plan-dumper/queries/tpch/tpch-q1.sql

@@ -1,3 +1,4 @@
+-- MODES: analyze, external-analyze
 -- This query is derived from TPC-H query 1
 -- THE TPC SOFTWARE IS AVAILABLE WITHOUT CHARGE FROM TPC.