Re:PostgreSQL 8.3 Released
by aantix on Monday February 04, 2008 @11:35AM
(#109146)
While I have enjoyed Postgres's reliability, the clients (for OSX) have left me feeling empty.
A majority of the time I just want to see the tables. When using pgAdmin3, I am forced to expand Servers->Local Server->Databases-><the desired database>->Schema->Public->Tables
And then when I finally get to the table I want, if I double click it, it brings up the properties for the table? Huh?
So I figure out I must click the table icon on the toolbar to bring up the table. Instead of bringing up the data within the same pane, it pops up a second window? Grrr....
While the tool gets the job done, I'm starting to miss MySQL just for the sake of having CocoaMySQL at my side.
Re:PostgreSQL 8.3 Released
by lneves on Monday February 04, 2008 @12:13PM
(#109161)
> When using pgAdmin3, I am forced to expand Servers->Local Server->Databases-><the desired database>->Schema->Public->Tables
Yes, that is annoying, but at least the 1.8.1 version of pgAdmin opens in the last db/schema used... that is the behavior I get.
> So I figure out I must click the table icon on the toolbar to bring up the table. Instead of bringing up the data within the same pane, it pops up a second window? Grrr....
File -> Options -> Preferences.
Unmark "Show object properties on doubleclick in treeview?"
Does that help?
Re:PostgreSQL 8.3 Released
by e1ven on Monday February 04, 2008 @10:48AM
(#109132)
Looking over the XML datatye [1], at first it doesn't seem that this would be very useful.. It doesn't seem to index the XML, only store it in a string-like manner..
But when I look at the XML functions [2], I can start to see where the real value is. In Chron X, we're storing vast sums of XML in SQL as strings, so being able to parse it per-element and pull them could be very helpful.
Re:PostgreSQL 8.3 Released
by inklesspen on Monday February 04, 2008 @11:52AM
(#109154)
It doesn't index the XML because it also doesn't provide any comparison methods for XML. Largely because string comparisons on XML tend to be less than useful, I think.
Re:PostgreSQL 8.3 Released
by bayareaguy on Tuesday February 05, 2008 @12:35AM
(#109546)
Although you can't index xml columns directly, you can define functional indexes over them. As long as you are careful to use the same function in your where clauses the index can be used.
Here's an example taken from Nikolay Samokhvalov's XML Support in PostgreSQL paper[1].
CREATE INDEX i_table1_xdata ON table1
USING btree( xpath_array(xdata, '//person/@name') );
Re:PostgreSQL 8.3 Released
by nradov on Monday February 04, 2008 @05:01PM
(#109320)
I'm not familiar with how they've done it in PostgreSQL, but DB2 does support XML indexes. You just have to write an XPath statement to represent what is stored in the index. Then you can write an XQuery statement and the query engine will use that index. It's quite fast.
Re:PostgreSQL 8.3 Released
by bayareaguy on Tuesday February 05, 2008 @12:47AM
(#109551)
Unfortunately the current PostgreSQL optimizer/executor doesn't do as much as it could when joining records from subtables partitioned with inherits clauses.
If you have a query with a predicate that can be pushed down to a subtable index then you're probably ok. However if you are doing something more intensive like an unqualified aggregation then you'll be in for a disappointment since the executor doesn't know how to preserve ordering when it scans the subtables. It will always give you a hash join or a merge join involving an intermediate sort even when the sort could be avoided by using an index scan of the subtables with something like a priority queue.
This can be worked around by doing your own incremental aggregations over the subtables and combining them, but it's a little bit of a pain.
A majority of the time I just want to see the tables. When using pgAdmin3, I am forced to expand Servers->Local Server->Databases-><the desired database>->Schema->Public->Tables
And then when I finally get to the table I want, if I double click it, it brings up the properties for the table? Huh?
So I figure out I must click the table icon on the toolbar to bring up the table. Instead of bringing up the data within the same pane, it pops up a second window? Grrr....
While the tool gets the job done, I'm starting to miss MySQL just for the sake of having CocoaMySQL at my side.
Yes, that is annoying, but at least the 1.8.1 version of pgAdmin opens in the last db/schema used... that is the behavior I get.
> So I figure out I must click the table icon on the toolbar to bring up the table. Instead of bringing up the data within the same pane, it pops up a second window? Grrr....
File -> Options -> Preferences. Unmark "Show object properties on doubleclick in treeview?" Does that help?
I use a combination of tools:
Aqua Data Studio
PgMaestro via Parallels
http://ourdoings.com/
POSTGRES AT HI5, Paul Lindner / Ram Gudavalli, June 12th, 2007
- http://www.inuus.com/talks/hi5_pug_06122007.pdf
http://www.massify.com
Harvard's Leadership for a Networked World uses PG - http://lnwprogram.org
DIYPlanner.com uses Drupal/PG - http://diyplanner.com
PowerfulIntentions.com (lots of forums, separate user pages, etc. all dynamically generated from PG database) - http://www.powerfulintentions.com
But when I look at the XML functions [2], I can start to see where the real value is. In Chron X, we're storing vast sums of XML in SQL as strings, so being able to parse it per-element and pull them could be very helpful.
[1]http://www.postgresql.org/docs/8.3/static/datatype-xml.html
[2]http://www.postgresql.org/docs/8.3/static/functions-xml.html
Here's an example taken from Nikolay Samokhvalov's XML Support in PostgreSQL paper[1].
[1] - http://www.pgcon.org/2007/schedule/attachments/16-xml_in_pos...http://enfranchisedmind.com/blog/2006/11/04/postgres-for-the...
Any comments?
If you have a query with a predicate that can be pushed down to a subtable index then you're probably ok. However if you are doing something more intensive like an unqualified aggregation then you'll be in for a disappointment since the executor doesn't know how to preserve ordering when it scans the subtables. It will always give you a hash join or a merge join involving an intermediate sort even when the sort could be avoided by using an index scan of the subtables with something like a priority queue.
This can be worked around by doing your own incremental aggregations over the subtables and combining them, but it's a little bit of a pain.
http://images.omniti.net/omniti.com/talks/partitions-public....