andygrove commented on PR #3783:
URL: 
https://github.com/apache/datafusion-comet/pull/3783#issuecomment-5441788095

   > **Note on this review:** this was generated by an LLM (Claude Code) at my 
request while I worked through a review backlog. I have not verified the 
individual findings myself. Please treat everything below as suggestions to 
evaluate rather than as authoritative review feedback, and push back on 
anything that is wrong or already handled.
   
   Adding a Spark-plus-Iceberg baseline engine so Comet's native Iceberg scan 
can be compared against the JVM Iceberg path is clearly useful, and per-query 
plan capture is the kind of thing you want the first time a benchmark result 
looks surprising.
   
   A few things.
   
   **The PR has been open a long time with an unaddressed review**
   
   There is review feedback on this from May. If those comments have been 
handled, a reply saying so would let it move; if they have not, that is the 
blocker rather than anything below.
   
   **Plan capture output**
   
   `tpcbench.py` gains per-query plan capture. Where do the plans go, how large 
do they get for TPC-DS at scale, and is the capture on by default or behind a 
flag? Twenty lines of diff is small, but writing a physical plan per query per 
engine can produce a lot of files, and a benchmark harness that quietly fills a 
directory is unpleasant. If it is opt-in, the description should say which flag.
   
   **Table creation accepting both shapes**
   
   Accepting `table.parquet` and `table/` is a two-line change and clearly 
convenient. Is there a case where both exist and the choice is ambiguous? If 
so, which wins, and should it warn?
   
   **Documentation**
   
   The description contains a good runbook: download Spark, fetch the Iceberg 
runtime jar, and so on. That is exactly the kind of thing that gets lost in a 
PR body. Is there a `benchmarks/` README where those steps belong, so someone 
who finds `spark-iceberg.toml` in six months knows how to use it?
   


-- 
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]

Reply via email to