search

prestodb / presto

The official home of the Presto distributed SQL query engine for big data
http://prestodb.io
Apache License 2.0
15.74k stars 5.28k forks source link

Add release notes for 0.288 #23079

Open wanglinsong opened 5 days ago

wanglinsong commented 5 days ago

Missing Release Notes

Ajay Gupte

Amit Dutta

Christian Zentgraf

Deepak Majeti

Emanuel F

Jialiang Tan

Jimmy Lu

Karteekmurthys

Ke

Konjac Huang

Linsong Wang

Sreeni Viswanadha

Vivek

Zac Blanco

jackychen718

wypb

xiaoxmeng

Extracted Release Notes

All Commits

yingsu00 commented 4 days ago

~~@gupteaj Ajay, BEFORE syntax support for presto engine.

22631 needs~~

1. Release note 2. The documentation in iceberg.rst.

I see they are added.

yingsu00 commented 4 days ago

@tanjialiang Somehow the release note for [[native] Remove spill stats reporting from presto native(https://github.com/prestodb/presto/pull/22751#top)#22751 was not picked up. But anyways your original release note might be too long. I 'd like to use the following:

Replace the presto_cpp.spill_metrics_name stats to velox.spill_metrics_name.

Do you think this is correct?

yingsu00 commented 4 days ago

@tangjiangling Can you please confirm and complete the following release note for [native] Migrate away stats reporting for allocator and cache ? Or you can provide a release note that follows https://github.com/prestodb/presto/wiki/Release-Notes-Guidelines.

Move the stats of memory allocator, in-memory cache and ssd cache from presto_cpp namespace to velox. For example, presto_cpp.memoryCacheStats.numEntries is now velox.memoryCacheStats.numEntries. The affected metrics are: ....
yingsu00 commented 4 days ago

@kewang1024 Can you please add a release note for "Fix hash calculation for Timestamp "? Thanks! Added

yingsu00 commented 4 days ago

@konjac-h Is there notable performance improvement to CTE selections? If yes, can you please provide a release note? Does the following sound good?

Improve the performance of reading common table expressions (CTE).

confirmed

konjac-h commented 4 days ago

@konjac-h Is there notable performance improvement to CTE selections? If yes, can you please provide a release note? Does the following sound good?

Improve the performance of reading common table expressions (CTE).

The fix reduces the unnecessary call to metastore API and improves the performance. Yes, the release note looks good to me.

wanglinsong commented 4 days ago

@yingsu00 please add review in the file directly, so that I know the location to update.

yingsu00 commented 2 days ago

@yingsu00 please add review in the file directly, so that I know the location to update.

I want to get the author's confirmation first. Then will add review.

yingsu00 commented 2 days ago

@ClarenceThreepwood The release note for Support for NOT NULL column constraints needs a bit touch up. Does it mean "NOT NULL" is not supported in the DDLs in general, but only take effects for Hive connector? What about Iceberg connector or Hudi? Could you please refine the release note to be more precise and easier to understand by the users?

ClarenceThreepwood commented 2 days ago

@ClarenceThreepwood The release note for Support for NOT NULL column constraints needs a bit touch up. Does it mean "NOT NULL" is not supported in the DDLs in general, but only take effects for Hive connector? What about Iceberg connector or Hudi? Could you please refine the release note to be more precise and easier to understand by the users?

It is not supported in iceberg or hudi yet. You can only use the vanilla hive connector to create constraints in HMS. I tweaked the note slightly to reflect this - let me know if you have a better rewrite in mind

yingsu00 commented 2 days ago

@xiaoxmeng Can you please confirm if the following release notes are correct? I made some changes to your PR message, but they still need to be clearer. For example you used "the minimum amount of memory", does it mean this amount is guaranteed to be reserved at least and it could be more? Would it be more precise just to say is this the total amount of memory reserved for xxx? You mentioned "A query can only allocate from the reserved query memory if its current capacity is less than the minimal memory capacity as specified by 'memory-pool-reserved-capacity'. Does it mean query-reserved-memory-gb has to be less than memory-pool-reserved-capacity?

Add two system configuration properties to specify the reserved query memory capacity on native clusters:
``query-reserved-memory-gb`` is the total amount of memory in gigabytes reserved for a
query. ``memory-pool-reserved-capacity`` is the total amount of memory in bytes reserved for each query memory pool.

cc @steveburnett

tdcmeehan commented 2 days ago

@yingsu00 can you please apply these suggestions to Linsong's branch? You can push directly to this branch.