Edition of Friday 13 November 2026

TParquet – Parquet for time-series, 50% compression with Pcodec

Comments on “TParquet – Parquet for time-series, 50% compression with Pcodec”

Claim-checked. Scored by all three judges, and its checkable claims were compared with the sources it cites. How far stories are checked

tparquet is an open-source extension to Parquet that adds a PCODEC compression codec for numeric columns, built on the pcodec library. On 151.5M rows of trade, quote, and order-book data from four market data providers, its files came out at 0.50× the size of ZSTD(9) Parquet, and every value reads…

Read the original at github.com. The judges saw only its text, with the site, the author and link domains removed; they are shown again after ranking.

Why it ranked

No. 22 of 137 with a score of 0.62. The panel rated it highest on new information (0.93) and technical depth (0.91), lowest on audience breadth (0.52). Placed No. 22, No. 57 by score alone: the placement terms moved other stories past it.

AI judgeJudge 2claude-haiku-5-5
A substantive, benchmarked release of a Parquet compression codec for numeric time series with reproducible benchmarks and specs; useful to data engineers, narrower in breadth, with some vendor-promotion tone.
AI judgeJudge 3gpt-5.6-luna
A technically detailed Parquet-compatible codec and tooling release reports reproducible benchmarks, cross-language integrations, conformance testing, and explicit compatibility trade-offs. It is highly actionable for data and market-data engineers, though adoption remains limited until broader reader support exists.

Each judge wrote its line after answering the rubric, with the site and the author still hidden. Every answer behind this is in the table under Score.

Score

0.62total under rubric v2.3

Judges’ mean answer per question, on a scale from 0 to 1. Questions marked minus count against the score.
  • Significance0.52
  • Reader interest0.72
  • Audience breadth0.52
  • Practical relevance0.76
  • Substantiated0.86
  • Technical depth0.91
  • New information0.93
  • Counts against: Promotional0.20
  • Counts against: Overclaim0.24

Each judge’s answer per question (0–1), their mean and spread, and what the question adds to the score (weight × mean).

QuestionJudge 1Judge 2Judge 3MeanSpreadWeightContribution (weight times mean)
SignificanceHow much does this change what people in tech know or can do?0.550.350.660.520.31+0.24+0.124
Reader interestWould a well-informed senior engineer or tech founder want to read this today, and likely pass it on?0.700.600.860.720.26+0.26+0.187
Audience breadthHow many people who build technology does it concern?0.520.400.630.520.23+0.13+0.067
Practical relevanceHow useful is it to someone who builds software or tech products?0.860.600.820.760.26+0.22+0.167
SubstantiatedHow well is it backed? Consider whether it is the primary source, and whether its main claims come with data, code, a demo, a reproducible method or citations.0.990.700.900.860.29+0.05+0.043
Technical depthHow technically substantive is it?0.990.800.940.910.19+0.05+0.046
New informationDoes it contain information absent from the stories covered in the last 7 days (list provided)?0.910.900.980.930.08+0.05+0.047
PromotionalIs it mainly marketing with little substance (a press release, ad or sponsored post with no real information)? A launch post that carries real news is not promotional.0.080.350.180.200.27−0.19−0.039
OverclaimDo the headline or conclusions claim more than the body supports?0.260.350.120.240.23−0.11−0.027
Sum of contributions+0.616
Claim-check penalty0.000
Total, as published0.616

Spread is the highest minus the lowest judge answer. marks a spread above 0.40: the judges disagree, and the story is marked contested.

The badge for this ranking, for the author's page or anyone's: one line of HTML, and it links back here.

Ranked #22 on Rubric, 13 Nov 2026
<a href="https://rubric.hrih.si/editions/2026-11-13/22/"><img src="https://rubric.hrih.si/badge/2026-11-13/22.svg" alt="Ranked #22 on Rubric, 13 Nov 2026" width="245" height="28"></a>

Claim check

Claim check: 5 matches source · 1 no source given · 2 couldn’t check. Each claim is compared with the source the article itself cites; a flag is published only when both checkers agree.

  1. Couldn’t checkBenchmark result

    tparquet compresses trade and tick files to about half their ZSTD size (0.49× on 139M trades from four market data providers), losslessly.

    The source was unreachable, paywalled or too long, or the two checkers disagreed.

    The article says

    “Trade and tick files at about half their ZSTD size (0.49× on 139M trades from four market data providers), lossless”

    The source says

    No passage could be compared.

    Source cited by the article: github.com/aperiodic-io/tparquet/blob/main/bench/RESULTS.md

  2. Matches sourceBenchmark result

    Across all 151.5M rows, tparquet files are 0.50× the size of ZSTD(9), 0.68× of tuned ZSTD(9) and 0.74× of the ALP baseline.

    The cited source says the same thing.

    The article says

    “All 151.5M 957.1 MB 700.7 MB 649.5 MB 478.6 MB 0.50× 0.68× 0.74×”

    The source says

    “All** | 31 | 151,499,256 | 957.1 MB | 700.7 MB | 649.5 MB | 478.6 MB | **0.500×** | **0.683×** | **0.737×”

    Source cited by the article: github.com/aperiodic-io/tparquet/blob/main/bench/RESULTS.md

  3. Matches sourceNumber

    tparquet is smaller than the baseline on every one of the 31 datasets, and 478 of the 480 numeric columns got PCODEC.

    The cited source says the same thing.

    The article says

    “tparquet is smaller on every one of the 31 datasets. It chooses PCODEC or ZSTD per column, on the first batch: here 478 of the 480 numeric columns got PCODEC”

    The source says

    “All** | 31 | 151,499,256 | 957.1 MB | 700.7 MB | 649.5 MB | 478.6 MB | **0.500×** | **0.683×** | **0.737×”

    Source cited by the article: github.com/aperiodic-io/tparquet/blob/main/bench/RESULTS.md

  4. Matches sourceNumber

    On one thread, tparquet writes at 102 MB/s of Arrow data and reads at 924 MB/s, versus ZSTD(9) at 97 MB/s write and 916 MB/s read.

    The cited source says the same thing.

    The article says

    “On one thread, tparquet writes at 102 MB/s of Arrow data (ZSTD(9): 97, tuned: 141, ALP: 169) and reads at 924 MB/s (ZSTD(9): 916, tuned: 1,129, ALP: 1,231”

    The source says

    “All** | 9.46 GB | 97 | 141 | 169 | 102 | 916 | 1,129 | 1,231 | 924”

    Source cited by the article: github.com/aperiodic-io/tparquet/blob/main/bench/RESULTS.md

  5. No source givenDate

    pcodec began as q_compress in May 2021, was redesigned as pcodec in 2023, and reached version 1.0 with a stable format in January 2026.

    The article gives no link or citation for it.

    The article says

    “It began as q_compress in May 2021 and was redesigned as pcodec in 2023. Version 1.0, with a stable format, came out in January 2026”

    The source says

    No source to compare with.

  6. Matches sourceNumber

    pcodec has 58 tagged releases and 19 contributors, and version 1.1.0 came out in November 2026.

    The cited source says the same thing.

    The article says

    “58 tagged releases and 19 contributors so far, on a regular release cadence. 1.1.0 came out in November 2026”

    The source says

    “v1.1.0 (v1.1.0, 2026-11-03”

    Source cited by the article: github.com/pcodec/pcodec

  7. Couldn’t checkGeneralization

    pcodec is described in a peer-reviewed paper and is a codec in Zarr and in the Vortex file format.

    The source was unreachable, paywalled or too long, or the two checkers disagreed.

    The article says

    “Peer-reviewed and widely used. It is described in the paper Pcodec: Better Compression for Numerical Sequences, and is a codec in Zarr (numcodecs, zarrs) and in the Vortex file format”

    The source says

    No passage could be compared.

    Source cited by the article: arxiv.org/abs/2502.06112

  8. Matches sourceCausal conclusion

    Vortex's default compressor leaves out heavyweight codecs such as pcodec to keep random access.

    The cited source says the same thing.

    The article says

    “Its default compressor leaves out heavyweight codecs such as pcodec to keep random access”

    The source says

    “the default compressor excludes opaque codecs like ZSTD and PCodec that sacrifice random access, while the Compact compressor includes them for maximum compression”

    Source cited by the article: vortex.dev/blog/btrblocks-compressor

Also covered by

No other coverage was clustered with this story.

Elsewhere

Elsewhere: HN 6 pts, not on front page

Discussion on Hacker News

Recorded for context only, checked 15 Nov 2026, 20:42 UTC. It played no part in the ranking.

Panel for this edition

Four AI agents ranked this edition: three judges scored every story blind to its site and author, and a fourth compared the top stories’ claims with the sources they cite. Summaries are written by AI.

Model versions are pinned for every edition. How ranking works describes rubric v2.3.

Comments

Readers’ comments on this story have a page of their own, ranked by a public rubric rather than by votes.

Comments on “TParquet – Parquet for time-series, 50% compression with Pcodec”

More on Rubric

From the last 7 editions, chosen by shared topic, score and date.

  1. Score 0.75

    llama.cpp v1.0