comphead commented on code in PR #214:
URL: https://github.com/apache/datafusion-site/pull/214#discussion_r4393589111
##########
content/blog/2026-10-25-comet-nested-tpch.md:
##########
@@ -0,0 +1,210 @@
+---
+layout: post
+title: Announcing a derived TPC-H, a benchmark for nested data in Apache
DataFusion Comet
+date: 2026-10-25
+author: pmc
+categories: [performance, benchmarks]
+---
+
+<!--
+{% comment %}
+Licensed to the Apache Software Foundation (ASF) under one or more
+contributor license agreements. See the NOTICE file distributed with
+this work for additional information regarding copyright ownership.
+The ASF licenses this file to you under the Apache License, Version 2.0
+(the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+{% endcomment %}
+-->
+
+[TOC]
+
+Query engines get fast at what their benchmarks measure, and the most widely
used analytical benchmarks contain no nested data. Apache DataFusion Comet now
runs all 22 queries of a nested TPC-H at scale factor 1000 (1 TB) in 777 s,
against 1,515 s for Spark. Before the fixes this benchmark exposed, it took
1,088 s.
+
+<img
+src="../images/comet-nested-tpch/intro-total.svg"
+width="100%"
+class="img-fluid"
+alt="Total execution time at SF1000: Spark 1,515.0 s, Comet nested unoptimized
1,087.5 s, Comet nested optimized 777.2 s"
+/>
+
+Nested TPC-H is a TPC-H derived benchmark in the Comet repository
([#7044](https://github.com/apache/datafusion-comet/pull/7044)). It keeps the
TPC-H data and questions and changes only the shape of the tables. The data and
queries are engine neutral, so they can be used to compare Spark, Comet and
other engines.
+
+We built it because our benchmarks and our users disagreed. Comet tracked its
performance with TPC-H and TPC-DS, and the numbers were encouraging. Real user
pipelines showed smaller gains, or none at all. What those pipelines had in
common was nested data, which neither benchmark contains.
+
+So we started experimenting with nested benchmarks in Comet to see where the
gaps were. The first runs showed them: some queries ran slower on Comet than on
Spark, even though Comet accelerates the same queries on flat TPC-H.
+
+## Why: TPC-H is flat, real data is not
+
+[TPC-H](https://www.tpc.org/tpch/) is one of the most cited analytical
benchmarks, and its schema is flat: eight normalized tables with primitive
columns. Its 22 queries exercise well-understood choke points such as
aggregation, join ordering and correlated subqueries (Boncz, Neumann and
Erling, "TPC-H Analyzed", TPCTC 2013).
+
+Real data is often nested. Event logs, JSON, Avro and Protobuf records, and
many data lake tables are full of structs and arrays, which Parquet stores with
the record shredding scheme from Dremel (Melnik et al., VLDB 2010). Nested data
adds work that flat benchmarks never exercise:
+
+- **Reading.** Rebuilding list offsets and validity bitmaps from repetition
and definition levels, and decoding only the fields a query uses.
+- **Planning.** Nested column pruning and predicate pushdown through `explode`.
+- **Operators.** `explode`, field access, and sort, join and exchange over
wide variable-size values.
+- **Memory and spilling.** Variable-size values make memory use and spill
volume harder to predict.
+- **Conversion.** Moving nested values between a JVM and a native
representation.
+
+Comet hit the shuffle case in
[#5397](https://github.com/apache/datafusion-comet/issues/5397), where native
shuffle writes of wide nested fields were two to three times slower than
Spark's. The issue notes that it was "invisible for TPC-* benchmarks because of
historical flat schema".
+
+We needed a benchmark that fills the gap between academic benchmarking and
real user workloads.
+
+## How: an adapter that nests TPC-H
+
+We did not want to invent a new benchmark from scratch. TPC-H is well
understood, widely published and easy to generate, and its business shape
(customers, orders, line items, suppliers and parts) is what makes its results
meaningful. The goal was to preserve that shape and those questions while
making every query process nested data.
+
+The idea turned out to be surprisingly simple: a transformation of both the
data and the queries that leaves the key and partition columns at the top level
and nests everything else. The partition columns are `p_brand`, `c_mktsegment`,
`o_orderdate` and `l_shipdate`, and every other non-key column moves into one
`array<struct>` named `<table>_data`. Each array holds exactly one element, so
exploding it never changes the row count and the queries return the same
answers as TPC-H. For `lineitem`:
+
+```
+-- flat -- nested
+l_orderkey BIGINT l_orderkey BIGINT
+l_partkey BIGINT l_partkey BIGINT
+l_suppkey BIGINT l_suppkey BIGINT
+l_linenumber INT l_shipdate DATE -- partition
column
+l_quantity DECIMAL(12,2) lineitem_data ARRAY<STRUCT<
+l_extendedprice DECIMAL(12,2) l_linenumber: INT,
+l_discount DECIMAL(12,2) l_quantity: DECIMAL(12,2),
+l_tax DECIMAL(12,2) l_extendedprice: DECIMAL(12,2),
+l_returnflag STRING l_discount: DECIMAL(12,2),
+l_linestatus STRING l_tax: DECIMAL(12,2),
+l_shipdate DATE l_returnflag: STRING,
+l_commitdate DATE l_linestatus: STRING,
+l_receiptdate DATE l_commitdate: DATE,
+l_shipinstruct STRING l_receiptdate: DATE,
+l_shipmode STRING l_shipinstruct: STRING,
+l_comment STRING l_shipmode: STRING,
+ l_comment: STRING>>
+```
+
+Each of the 22 queries is rewritten to unnest only the arrays it needs. The
queries are in
[`benchmarks/nested/queries`](https://github.com/apache/datafusion-comet/tree/main/benchmarks/nested/queries).
+
+```sql
+from (select l_orderkey, l_partkey, l_suppkey, l_shipdate,
+ explode(lineitem_data) as lineitem_data from lineitem) as lineitem
+where lineitem_data.l_quantity < 24
+```
+
+[`nest_tpch.scala`](https://github.com/apache/datafusion-comet/blob/main/benchmarks/nested/scripts/nest_tpch.scala)
converts an existing flat TPC-H dataset to the nested layout. The source and
the target can each be Parquet or Iceberg. The script runs under `spark-shell`
on a laptop or a cluster, casts columns to fixed types, writes zstd-compressed
Parquet with 512 MiB row groups, partitions by the columns above and checks row
counts:
+
+```bash
+spark-shell -i scripts/nest_tpch.scala \
+ --conf spark.nesttpch.source=parquet:/data/tpch/sf10 \
+ --conf spark.nesttpch.target=parquet:/data/tpch-nested/sf10
+```
+
+The
[README](https://github.com/apache/datafusion-comet/tree/main/benchmarks/nested)
has the Iceberg variants and all settings.
+
+## Benchmark setup
+
+- **Data**: nested TPC-H, scale factor 1000 (1 TB).
+- **Tables**: Apache Iceberg tables with zstd-compressed Parquet data files on
S3, created with `nest_tpch.scala`. Comet reads them with its native Iceberg
scan, which is built on [iceberg-rust](https://github.com/apache/iceberg-rust).
+- **Cluster**: Kubernetes with 16-core executors.
+- **Method**: one run per configuration, three iterations per query. Figures
are the fastest iteration, and totals are sums of the per-query times.
+- **Spark**: Spark 4.0
+- **Comet nested unoptimized**: Comet 1.1.0 snapshot (revision `a50fe3aea863`)
+- **Comet nested optimized**: Comet 1.3.0 snapshot (revision `330b29cf`)
+
+## What nested benchmarks revealed
+
+The nested benchmark exposed issues that were invisible on flat TPC-H. In the
unoptimized run, Comet was slower than Spark on q20 (32.9 s against 31.5 s) and
q21 (490.6 s against 373.5 s), and barely faster on q02 (34.0 s against 34.9
s). Tracing those queries surfaced a lot of technical debt in how nested data
is read, shuffled and spilled, in Comet and in the iceberg-rust reader
underneath it.
+
+The issues fell into four groups:
+
+- **Spilling.** Wide nested values made q21's sort spill several hundred GiB
per iteration, and Comet wrote spill files uncompressed.
+- **Reading only what a query needs.** Comet's native Iceberg scan decoded
every field of a nested column, even when Spark had pruned it to one field.
This is why q20 was no faster than Spark.
+- **Shuffle.** Native shuffle wrote wide nested fields two to three times
slower than Spark.
+- **Nested reads in iceberg-rust.** Nested fields read as NULL, some list
types could not be read at all, and projection cost grew with the number of
nested leaves.
+
+Spills and shuffle writes happen in flat workloads too. The nested schema
changed the data volume or access pattern enough to expose them.
+
+Most of these issues are fixed. The issues and pull requests, with their
measurements, are listed in [Issues and pull
requests](#issues-and-pull-requests) at the end of the post.
+
+## Results
+
+### Total execution time
+
+Unoptimized Comet takes `1,087.5` s against `1,515.0` s for Spark (1.39x) and
is slower on q20 and q21. Optimized Comet takes `777.2` s (1.95x), 29% less
than unoptimized, and is faster than Spark on all 22 queries.
+
+<img
+src="../images/comet-nested-tpch/intro-total.svg"
Review Comment:
its repeating yes, the idea was to keep the first chart as the teaser, and
results section replicates the same chart to keep all the results in a single
section and avoid the user to scroll back to the beginning for checking the
totals. WDYT? is it a good practice to repeat, or it should be structured some
other way?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
