Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

PostgreSQL 9.3 Beta 1 Released

postgresql.org

Post reply on HN
Print view View original

101–110 of 118 posts

Re: PostgreSQL 9.3 Beta 1 Released

#101

Post by fusiongyro »

pjungwir wrote: post #53
jeffdavis wrote: post #48

So, what's still missing in postgres? To make the responses as constructive as possible, please also specify: * Whether you currently use postgres, and what other systems you use (presumably ones that do have the feature that is missing in postgres) * Whether you currently need the feature, or whether you anticipate the need in the future, or whether you expect that other people will need it * Whether the missing fea…

I'd love to see on-disk bitmap indexes. I noticed there was an attempt to add these around 2009 but it was abandoned after being (apparently) nearly completed. [1,2] I'd like to see these because it would make Postgres better at data warehouse applications. I've been thinking about volunteering to add them myself, but I've never looked at the Postgres codebase before, so it seems a bit ambitious. A minor feature I'd…

> I'd love to see on-disk bitmap indexes.

I'm curious about what the performance ramifications of that would look like. I recently had to disable bitmap scans in a data warehouse-type-thing I run because filesystem fragmentation was apparently misleading Postgres into thinking it would be cheaper than an index scan. The bitmap scan plan takes ~15 minutes to run cold, and the index scan takes ~20 seconds cold. (Really odd thing is that index scan takes ~1.5 seconds warm, versus ~0.4 seconds warm with the bitmap scan).

Re: PostgreSQL 9.3 Beta 1 Released

#102

Post by ef4 »

joevandyk wrote: post #70
jeffdavis wrote: post #48

So, what's still missing in postgres? To make the responses as constructive as possible, please also specify: * Whether you currently use postgres, and what other systems you use (presumably ones that do have the feature that is missing in postgres) * Whether you currently need the feature, or whether you anticipate the need in the future, or whether you expect that other people will need it * Whether the missing fea…

I could really use UPSERT. I'm using postgresql right now. My use case: I could have a table like: create table items (id uuid primary key, attributes json); I'd have a webapp using javascript for adding/editing new items. If the user wants to add a new item, the client code would generate the uuid. Editing and adding new rows would both use the same UPSERT/MERGE code -- no need for separate insert or update statemen…

Upsert is already possible in a single statement, though a bit unwieldy.

    create table people (uuid varchar, name varchar);
    with new_rows as (select (unnest(ARRAY[('xxx','John Smith'), ('yyy', 'Ben Bitdiddle')]::people[])).*), updated as (UPDATE people SET name=new_rows.name FROM new_rows WHERE people.uuid=new_rows.uuid RETURNING people.uuid) insert into people select new_rows.* from new_rows left join updated using (uuid) where updated.uuid is null;

Re: PostgreSQL 9.3 Beta 1 Released

#103

Post by mamcx »

jeffdavis wrote: post #48

So, what's still missing in postgres? To make the responses as constructive as possible, please also specify: * Whether you currently use postgres, and what other systems you use (presumably ones that do have the feature that is missing in postgres) * Whether you currently need the feature, or whether you anticipate the need in the future, or whether you expect that other people will need it * Whether the missing fea…

No missing per-se, but I wish I could use postgres as a client db in iOS (like sqlite) so I can have a single codebase (like is possible with firebird) even if that mean only a subset (what have sqlite, plus psql store procedures!)

Re: PostgreSQL 9.3 Beta 1 Released

#104

Post by psadauskas »

joevandyk wrote: post #70
jeffdavis wrote: post #48

So, what's still missing in postgres? To make the responses as constructive as possible, please also specify: * Whether you currently use postgres, and what other systems you use (presumably ones that do have the feature that is missing in postgres) * Whether you currently need the feature, or whether you anticipate the need in the future, or whether you expect that other people will need it * Whether the missing fea…

I could really use UPSERT. I'm using postgresql right now. My use case: I could have a table like: create table items (id uuid primary key, attributes json); I'd have a webapp using javascript for adding/editing new items. If the user wants to add a new item, the client code would generate the uuid. Editing and adding new rows would both use the same UPSERT/MERGE code -- no need for separate insert or update statemen…

Here's an UPSERT in postgresql: https://gist.github.com/paul/855efdecaaa2ec4deec7

While we're at it, here's a "find or create" in a single atomic statement: https://gist.github.com/paul/75ec84d131e36492b17b

Re: PostgreSQL 9.3 Beta 1 Released

#105

Post by snaky »

fpic wrote: post #98
jeffdavis wrote: post #48

So, what's still missing in postgres? To make the responses as constructive as possible, please also specify: * Whether you currently use postgres, and what other systems you use (presumably ones that do have the feature that is missing in postgres) * Whether you currently need the feature, or whether you anticipate the need in the future, or whether you expect that other people will need it * Whether the missing fea…

HTTP APIs like http://wiki.postgresql.org/wiki/HTTP_API . Using a javascript framework like AngularJs would permit me to get rid of middleware (Java, PHP, dotNET, etc.) almost completely. I need a relational db and I'm not a big fan of NoSQL, so MongoDB or others are not a viable option. It would also permit to use only javascript from the front end to the back end ( https://code.google.com/p/plv8js/ ).

Nginx + ngx_postgres + ngx_rds_json (or OpenResty bundle) is perfect for the task. See sample #5 at https://github.com/FRiCKLE/ngx_postgres

And it's fast.

Re: PostgreSQL 9.3 Beta 1 Released

#106

Post by tibbon »

jeffdavis wrote: post #48

So, what's still missing in postgres? To make the responses as constructive as possible, please also specify: * Whether you currently use postgres, and what other systems you use (presumably ones that do have the feature that is missing in postgres) * Whether you currently need the feature, or whether you anticipate the need in the future, or whether you expect that other people will need it * Whether the missing fea…

I wish there were more built in replication options. Setting up Master-Master (data existing in multiple data centers) with hot slaves for example- I don't think you can do this with Postgres alone currently. Need to use something like Bucardo (a perl program) to make it happen.

Re: PostgreSQL 9.3 Beta 1 Released

#107

Post by daigoba66 »

jkbyc wrote: post #59
Refefer wrote: post #20

Materialized Views! Well, sort of; it appears they don't automatically refresh when data has been updated in the underlying tables ala Oracle's implementation, but this gets you rather close.

Implementing efficient updates of materialized views is difficult. They have to support recursive queries, aggregation, and probably even recursive views that they list as new in "Additional Features." Algorithms for updates of materialized views exist since at least 20 years but maybe they don't cover all these cases efficiently.

The implementation used by MSSQL is pretty restrictive in terms of what kind of query can be indexed. But it works rather well.

Re: PostgreSQL 9.3 Beta 1 Released

#108

Post by snaky »

twerquie wrote: post #52
jeffdavis wrote: post #48

So, what's still missing in postgres? To make the responses as constructive as possible, please also specify: * Whether you currently use postgres, and what other systems you use (presumably ones that do have the feature that is missing in postgres) * Whether you currently need the feature, or whether you anticipate the need in the future, or whether you expect that other people will need it * Whether the missing fea…

Multi-master in the core distribution.

Postgres-XC?

Re: PostgreSQL 9.3 Beta 1 Released

#109

Post by jap »

jeffdavis wrote: post #48

So, what's still missing in postgres? To make the responses as constructive as possible, please also specify: * Whether you currently use postgres, and what other systems you use (presumably ones that do have the feature that is missing in postgres) * Whether you currently need the feature, or whether you anticipate the need in the future, or whether you expect that other people will need it * Whether the missing fea…

Long time PostgreSQL user. Would like to use table inheritance in a couple places, but don't because of these caveats: http://www.postgresql.org/docs/9.2/static/ddl-inherit.html#D...

Re: PostgreSQL 9.3 Beta 1 Released

#110

Post by ocharles »

psadauskas wrote: post #104
joevandyk wrote: post #70

Earlier quoted context omitted.

I could really use UPSERT. I'm using postgresql right now. My use case: I could have a table like: create table items (id uuid primary key, attributes json); I'd have a webapp using javascript for adding/editing new items. If the user wants to add a new item, the client code would generate the uuid. Editing and adding new rows would both use the same UPSERT/MERGE code -- no need for separate insert or update statemen…

Here's an UPSERT in postgresql: https://gist.github.com/paul/855efdecaaa2ec4deec7 While we're at it, here's a "find or create" in a single atomic statement: https://gist.github.com/paul/75ec84d131e36492b17b

I believe both of these still have race conditions, and are not the same as truly atomic upserts. Granted the window is considerably smaller, but their is still a race condition.
Post reply on HN