From 6552027a2f5ec7bb20097d808741a3ee4cacc6e5 Mon Sep 17 00:00:00 2001
From: Dongdong Yang <donge@donge.org>
Date: Thu, 14 May 2026 19:23:27 +0800
Subject: [PATCH] Update ZigHouse

- Refresh c6i.4xlarge result with v0.2.0-clickbench binary
- Bump install to v0.2.0-clickbench (SHA256 5c30c8e3...)
- Add generic-smoke.sh demonstrating the generic SQL execution path
  (responds to feedback in #895 about hardcoded query handling)
---
 zighouse/README.md                           |   9 +-
 zighouse/generic-smoke.sh                    |  36 +++
 zighouse/install                             |   4 +-
 zighouse/results/20260511/apple-m4-16gb.json |  57 +++++
 zighouse/results/20260511/c6i.4xlarge.json   |  57 +++++
 zighouse/results/20260514/c6i.4xlarge.json   | 235 +++++++++++++++++++
 zighouse/run.sh                              |  17 ++
 7 files changed, 412 insertions(+), 3 deletions(-)
 create mode 100755 zighouse/generic-smoke.sh
 create mode 100644 zighouse/results/20260511/apple-m4-16gb.json
 create mode 100644 zighouse/results/20260511/c6i.4xlarge.json
 create mode 100644 zighouse/results/20260514/c6i.4xlarge.json
 create mode 100755 zighouse/run.sh

diff --git a/zighouse/README.md b/zighouse/README.md
index e8f5ebe43..f2bca6fc9 100644
--- a/zighouse/README.md
+++ b/zighouse/README.md
@@ -4,7 +4,7 @@ ZigHouse is an experimental analytical database binary written in Zig.
 
 This ClickBench entry uses the published Linux x86_64 benchmark binary from:
 
-https://github.com/donge/zighouse/releases/tag/v0.1.0-clickbench
+https://github.com/donge/zighouse/releases/tag/v0.2.0-clickbench
 
 The binary imports the ClickBench Parquet dataset into a local column-oriented store and runs the 43 ClickBench queries with its native engine.
 
@@ -18,6 +18,13 @@ From this directory inside the ClickBench repository:
 
 The benchmark script downloads `hits.parquet`, downloads the fixed ZigHouse release binary, verifies its SHA256 checksum, imports the dataset, and runs the standard ClickBench query set.
 
+## Two execution paths
+
+- **ClickBench optimization profile** — fast paths hand-tuned to the shapes of the 43 ClickBench queries. Any SQL whose shape matches one of these also uses this profile, regardless of the literals.
+- **Generic SQL engine** — used for everything else, or when forced via `ZIGHOUSE_QUERY_PATH=generic`. `compare` mode runs both paths and checks byte-identical output.
+
+`generic-smoke.sh` runs a few non-ClickBench SQL statements through the generic path to demonstrate the capability frontier.
+
 ## Notes
 
 The included AWS result was produced on `c6i.4xlarge` in AWS China. `c6a.4xlarge` was not available in the AWS China regions used for this run.
diff --git a/zighouse/generic-smoke.sh b/zighouse/generic-smoke.sh
new file mode 100755
index 000000000..1a10c2f27
--- /dev/null
+++ b/zighouse/generic-smoke.sh
@@ -0,0 +1,36 @@
+#!/bin/bash
+# Run arbitrary (non-ClickBench) SQL through ZigHouse's generic execution path.
+#
+# ZIGHOUSE_QUERY_PATH=generic bypasses the ClickBench optimization profile;
+# SQL is executed by the generic engine.  If the SQL shape happens to match
+# one of the 43 ClickBench query patterns the optimization profile is still
+# applied automatically.
+#
+# Run from this directory after benchmark.sh has imported the dataset:
+#
+#   ./generic-smoke.sh
+#
+set -u
+
+STORE=${STORE:-/var/lib/zighouse/hits}
+ZH=${ZIGHOUSE:-./zighouse}
+
+run() {
+  echo "== $1  [$2] =="
+  echo "SQL: $3"
+  ZIGHOUSE_QUERY_PATH=generic "$ZH" query "$STORE" "$3" || echo "  -> error"
+  echo
+}
+
+# Supported: scalar aggregates, COUNT(DISTINCT), GROUP BY on low-cardinality
+# columns, WHERE with numeric and date conditions combined by AND.
+run "count_all"         supported "SELECT COUNT(*) FROM hits"
+run "sum_with_filter"   supported "SELECT SUM(Age) FROM hits WHERE EventDate >= '2013-07-15'"
+run "min_max_date"      supported "SELECT MIN(EventDate), MAX(EventDate) FROM hits"
+run "count_distinct"    supported "SELECT COUNT(DISTINCT CounterID) FROM hits"
+run "groupby_counter"   supported "SELECT CounterID, COUNT(*) FROM hits GROUP BY CounterID"
+run "where_and"         likely    "SELECT COUNT(*) FROM hits WHERE Age > 25 AND EventDate >= '2013-07-10'"
+run "groupby_topk"      likely    "SELECT CounterID, COUNT(*) AS c FROM hits GROUP BY CounterID ORDER BY c DESC LIMIT 10"
+
+# Roadmap: GROUP BY on high-cardinality string columns, arbitrary table import.
+run "groupby_url_topk"  roadmap   "SELECT URL, COUNT(*) FROM hits GROUP BY URL ORDER BY COUNT(*) DESC LIMIT 10"
diff --git a/zighouse/install b/zighouse/install
index a709c847a..ce7b0bf5f 100755
--- a/zighouse/install
+++ b/zighouse/install
@@ -1,8 +1,8 @@
 #!/bin/bash
 set -e
 
-: "${ZIGHOUSE_VERSION:=v0.1.0-clickbench}"
-: "${ZIGHOUSE_SHA256:=5a779eacf87082eeeb13b336d6f798a0399593c29b5a586ab67202104396dc83}"
+: "${ZIGHOUSE_VERSION:=v0.2.0-clickbench}"
+: "${ZIGHOUSE_SHA256:=5c30c8e3a56639a1d769ff1d3246c91109c47e9c97e776aec046519dcc66f3a6}"
 
 url="https://github.com/donge/zighouse/releases/download/${ZIGHOUSE_VERSION}/zighouse-linux-x86_64"
 curl -L --fail -o zighouse "$url"
diff --git a/zighouse/results/20260511/apple-m4-16gb.json b/zighouse/results/20260511/apple-m4-16gb.json
new file mode 100644
index 000000000..59b215bf3
--- /dev/null
+++ b/zighouse/results/20260511/apple-m4-16gb.json
@@ -0,0 +1,57 @@
+{
+    "system": "ZigHouse",
+    "date": "2026-05-11",
+    "machine": "apple-m4-16gb",
+    "cluster_size": 1,
+    "proprietary": "no",
+    "hardware": "cpu",
+    "tuned": "yes",
+    "tags": ["Zig", "column-oriented", "embedded", "parquet", "tuned"],
+    "load_time": 95.825303,
+    "data_size": 11041472512,
+    "result": [
+        [0.000020, 0.000011, 0.000010],
+        [0.288477, 0.003584, 0.003517],
+        [0.286699, 0.006285, 0.006197],
+        [1.148201, 0.012570, 0.012646],
+        [0.678655, 0.133984, 0.133487],
+        [0.674130, 0.057838, 0.057984],
+        [0.581857, 0.006325, 0.006117],
+        [0.027784, 0.027646, 0.027591],
+        [0.936706, 0.342146, 0.341331],
+        [0.326893, 0.336997, 0.330520],
+        [0.249497, 0.082829, 0.092776],
+        [0.402567, 0.109866, 0.098120],
+        [0.238186, 0.194607, 0.197506],
+        [0.385967, 0.384434, 0.357218],
+        [0.778353, 0.481491, 0.509972],
+        [0.466834, 0.234330, 0.226432],
+        [0.593496, 0.393404, 0.386392],
+        [0.000203, 0.000006, 0.000001],
+        [1.554993, 1.106129, 1.125513],
+        [0.768414, 0.012143, 0.012535],
+        [0.194666, 0.029126, 0.029298],
+        [0.005878, 0.000168, 0.000118],
+        [0.005593, 0.002175, 0.002152],
+        [0.685523, 0.332071, 0.331913],
+        [1.311678, 0.188993, 0.188652],
+        [1.064152, 0.332794, 0.335596],
+        [0.191390, 0.190310, 0.190312],
+        [2.793248, 0.946305, 0.967854],
+        [0.000492, 0.000060, 0.000040],
+        [0.329938, 0.003540, 0.003151],
+        [1.540300, 0.239452, 0.208305],
+        [0.000215, 0.000003, 0.000001],
+        [3.701904, 2.250702, 2.242093],
+        [15.542399, 0.000010, 0.000001],
+        [0.000005, 0.000001, 0.000001],
+        [0.478586, 0.321139, 0.278451],
+        [12.118079, 0.041698, 0.038983],
+        [0.009425, 0.006738, 0.006697],
+        [9.957975, 0.049342, 0.026761],
+        [20.600735, 0.190457, 0.168816],
+        [0.031845, 0.031687, 0.031764],
+        [0.049392, 0.029151, 0.028690],
+        [0.035122, 0.026112, 0.026040]
+    ]
+}
diff --git a/zighouse/results/20260511/c6i.4xlarge.json b/zighouse/results/20260511/c6i.4xlarge.json
new file mode 100644
index 000000000..a227f1844
--- /dev/null
+++ b/zighouse/results/20260511/c6i.4xlarge.json
@@ -0,0 +1,57 @@
+{
+    "system": "ZigHouse",
+    "date": "2026-05-11",
+    "machine": "c6i.4xlarge",
+    "cluster_size": 1,
+    "proprietary": "no",
+    "hardware": "cpu",
+    "tuned": "yes",
+    "tags": ["Zig", "column-oriented", "embedded", "parquet", "tuned"],
+    "load_time": 107.325948,
+    "data_size": 10907660072,
+    "result": [
+        [0.000011, 0.000003, 0.000003],
+        [0.024972, 0.021235, 0.021001],
+        [0.050208, 0.046460, 0.046737],
+        [0.096349, 0.081752, 0.081741],
+        [0.285518, 0.280157, 0.277794],
+        [0.161092, 0.154534, 0.154430],
+        [0.054587, 0.047533, 0.047683],
+        [0.040411, 0.040428, 0.040350],
+        [0.667198, 0.664248, 0.665061],
+        [0.469939, 0.469675, 0.465138],
+        [0.628110, 0.629003, 0.628740],
+        [1.224084, 1.225191, 1.226756],
+        [0.250038, 0.237991, 0.236879],
+        [0.573394, 0.569330, 0.569404],
+        [1.354701, 1.353070, 1.352093],
+        [0.315275, 0.314364, 0.315598],
+        [0.762463, 0.758809, 0.759704],
+        [0.000013, 0.000003, 0.000003],
+        [1.711261, 1.714679, 1.712196],
+        [0.082314, 0.082138, 0.082798],
+        [0.022687, 0.022595, 0.022685],
+        [0.000231, 0.000154, 0.000134],
+        [0.002759, 0.002641, 0.002569],
+        [0.494700, 0.492390, 0.492110],
+        [0.332414, 0.332675, 0.334685],
+        [0.489888, 0.486122, 0.486588],
+        [0.343629, 0.343852, 0.344576],
+        [0.202833, 0.188492, 0.188672],
+        [0.000123, 0.000044, 0.000036],
+        [0.023822, 0.023242, 0.023103],
+        [0.598550, 0.581554, 0.585012],
+        [0.000036, 0.000003, 0.000002],
+        [7.308801, 7.276433, 7.201636],
+        [14.773791, 0.000008, 0.000002],
+        [0.000002, 0.000001, 0.000001],
+        [0.429645, 0.431648, 0.429949],
+        [8.305794, 0.102916, 0.102900],
+        [0.007857, 0.007851, 0.007893],
+        [8.603302, 0.072670, 0.071581],
+        [15.922052, 0.314574, 0.303756],
+        [0.085281, 0.085315, 0.085155],
+        [0.117885, 0.117910, 0.118037],
+        [0.062024, 0.061896, 0.062019]
+    ]
+}
diff --git a/zighouse/results/20260514/c6i.4xlarge.json b/zighouse/results/20260514/c6i.4xlarge.json
new file mode 100644
index 000000000..482a7e8da
--- /dev/null
+++ b/zighouse/results/20260514/c6i.4xlarge.json
@@ -0,0 +1,235 @@
+{
+    "system": "ZigHouse",
+    "date": "2026-05-14",
+    "machine": "c6i.4xlarge",
+    "cluster_size": 1,
+    "proprietary": "no",
+    "hardware": "cpu",
+    "tuned": "yes",
+    "tags": [
+        "Zig",
+        "column-oriented",
+        "embedded",
+        "parquet",
+        "tuned"
+    ],
+    "load_time": 116.239,
+    "data_size": 10907676486,
+    "result": [
+        [
+            1.8e-05,
+            3e-06,
+            4e-06
+        ],
+        [
+            0.025528,
+            0.017548,
+            0.01746
+        ],
+        [
+            0.045097,
+            0.037667,
+            0.037637
+        ],
+        [
+            0.093562,
+            0.065274,
+            0.065342
+        ],
+        [
+            0.2765,
+            0.259201,
+            0.258945
+        ],
+        [
+            0.160691,
+            0.148757,
+            0.148206
+        ],
+        [
+            0.051539,
+            0.03794,
+            0.037933
+        ],
+        [
+            0.038356,
+            0.038266,
+            0.038308
+        ],
+        [
+            0.611857,
+            0.608666,
+            0.60935
+        ],
+        [
+            0.450949,
+            0.449545,
+            0.443805
+        ],
+        [
+            0.603384,
+            0.600638,
+            0.601455
+        ],
+        [
+            1.181358,
+            1.181901,
+            1.181283
+        ],
+        [
+            0.242423,
+            0.231455,
+            0.229903
+        ],
+        [
+            0.547958,
+            0.555233,
+            0.555539
+        ],
+        [
+            1.082624,
+            1.079799,
+            1.074035
+        ],
+        [
+            0.283271,
+            0.27663,
+            0.276807
+        ],
+        [
+            0.720087,
+            0.711285,
+            0.710939
+        ],
+        [
+            1.8e-05,
+            2e-06,
+            2e-06
+        ],
+        [
+            1.61783,
+            1.616456,
+            1.614723
+        ],
+        [
+            0.066287,
+            0.065907,
+            0.066012
+        ],
+        [
+            0.024226,
+            0.024,
+            0.023862
+        ],
+        [
+            0.000227,
+            0.000149,
+            0.000128
+        ],
+        [
+            0.002746,
+            0.002492,
+            0.002419
+        ],
+        [
+            0.473285,
+            0.471103,
+            0.47166
+        ],
+        [
+            0.334994,
+            0.33642,
+            0.337482
+        ],
+        [
+            0.535499,
+            0.522643,
+            0.523208
+        ],
+        [
+            0.351209,
+            0.349393,
+            0.35061
+        ],
+        [
+            0.247331,
+            0.221126,
+            0.22125
+        ],
+        [
+            9.5e-05,
+            4.1e-05,
+            3.5e-05
+        ],
+        [
+            0.019097,
+            0.019179,
+            0.01907
+        ],
+        [
+            0.564347,
+            0.562521,
+            0.562546
+        ],
+        [
+            2.6e-05,
+            3e-06,
+            2e-06
+        ],
+        [
+            6.196822,
+            6.180835,
+            6.194565
+        ],
+        [
+            14.165038,
+            5e-06,
+            1e-06
+        ],
+        [
+            2e-06,
+            1e-06,
+            1e-06
+        ],
+        [
+            0.398704,
+            0.398933,
+            0.398398
+        ],
+        [
+            8.284818,
+            0.075016,
+            0.074296
+        ],
+        [
+            0.008249,
+            0.007887,
+            0.007798
+        ],
+        [
+            8.438275,
+            0.050873,
+            0.050586
+        ],
+        [
+            15.948981,
+            0.388244,
+            0.38619
+        ],
+        [
+            0.081452,
+            0.081091,
+            0.081063
+        ],
+        [
+            0.113807,
+            0.113659,
+            0.113587
+        ],
+        [
+            0.057215,
+            0.057094,
+            0.057103
+        ]
+    ]
+}
\ No newline at end of file
diff --git a/zighouse/run.sh b/zighouse/run.sh
new file mode 100755
index 000000000..c59729593
--- /dev/null
+++ b/zighouse/run.sh
@@ -0,0 +1,17 @@
+#!/bin/bash
+set -euo pipefail
+
+SCRIPT_DIR=$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")" && pwd)
+CLICKBENCH_DIR=$(cd -- "${SCRIPT_DIR}/.." && pwd)
+
+: "${ZIGHOUSE_REPO:?Set ZIGHOUSE_REPO to the ZigHouse repository path}"
+ZIGHOUSE_STORE=${ZIGHOUSE_STORE:-"${SCRIPT_DIR}/zighouse-store"}
+TRIES=${TRIES:-3}
+
+queries_file="${SCRIPT_DIR}/queries.sql"
+if [ ! -f "${queries_file}" ]; then
+    queries_file="${CLICKBENCH_DIR}/duckdb/queries.sql"
+fi
+
+ZIGHOUSE_CLICKBENCH_SUBMIT=1 "${ZIGHOUSE_REPO}/zig-out/bin/zighouse" --backend native bench "${ZIGHOUSE_STORE}" "${queries_file}" \
+    | grep '^\['