#POSTGIS IRC Log - 2009-06-16

For logs after Feb 3, 2007, all times are GMT-8. Prior logs are GMT-9.
Back to Logs
02:06:23 simplexio: ha1331: yes.
02:08:07 simplexio: ha1331: use metric srid, or create line between two points and st_legth( st_transform( thatline, in metric srid ) )... or if irecall right there may be functio which return it straight.. it quite new one..
06:43:43 ltcabral: is there a query to retrieve the type of the geometric structure? if a polygon or a multiline or something..?
06:51:33 ue: ST_GeometryType
06:53:29 ltcabral: thanks
07:26:28 bitner: ha1331: simplexio more simply st_distance( ) for data in metric coords or st_distance_sphere( ) or st_distance_spheroid( ) for geographic coords
07:28:08 strk: uhm
07:29:17 CIA-34: strk * r2589 /trunk/ ( 6 files in 3 dirs ): Revert single-sided patch. Martin Davis is workin on it from the JTS side
07:38:46 CIA-34: strk * r2590 /trunk/source/ ( 2 files in 2 dirs ): Port info for RobustDeterminant
07:39:55 jalmeida: can anybody help me, i create a tool to merge polygons but i didn't have the best algoritm to merge polygons in 2d?
07:39:56 jalmeida: exist any good algoritm to merge poligons?
07:41:40 CIA-34: strk * r2591 /trunk/source/ ( 2 files in 2 dirs ): Sync RobustDeterminant to revision 1.14
07:45:23 CIA-34: strk * r2592 /trunk/source/ ( 2 files in 2 dirs ): Sync RobustDeterminant to JTS-1.10 ( r1.15 was just commented-out code )
07:47:53 ab0oo: can someone help me with geomFromText syntax? Specifically, I have a postgres function that, given a center point and "radius", the function will return a hexagon shaped polygon. But, I'm having a hard time giving variables to geomFromText. Anyone ever done this?
07:48:41 CIA-34: robe * r4175 /trunk/NEWS: Forgot ST_MinimumBoundingCircle and credit Bruce Rindahl
07:49:49 strk: geomFromText takes text
07:49:58 strk: you can use concatenation operator to intermix variables
07:50:14 strk: 'My ' || fruit || ' is ' || color
07:50:33 strk: jalmeida: GeomUnion "merges" polygons
07:51:03 strk: or probably st_Union is the prefer name today
07:51:13 : * strk has to admit the st_ prefix in this case did some good : )
07:51:48 ab0oo: strk: ty
07:53:18 bitner: st_rk: using st_ rather than the "simple" names is a tough habit to break
07:54:31 nhv: background on 'current' ST_Union http://blog.cleverelephant.ca/search?updated-max=2009-01-28T09%3A56%3A00-08%3A00&max-results=20
07:54:32 sigq: Title: Paul Ramsey ( at blog.cleverelephant.ca )
07:54:36 strk: the worst is you need two hands for a _ : )
07:56:01 : * bitner wonders how many things of his will break if/when the old names go away
08:05:27 ab0oo: strk: thank you again. That was the most hideous, painful, ugly text concatenation I've ever created, and it works perfectly.
08:07:14 strk: np
09:25:54 ha1331: I have openstreetmap data on my database, is it possible to get the street closest to a point?
09:27:21 ha1331: I'm trying to make tracking application and have tracking data from gps devices and would like to get the streetname for the points
11:05:17 pimpaa: hey guys
11:05:22 pimpaa: can any of you see a problem with this?
11:05:23 pimpaa: http://pgsql.privatepaste.com/cegAPPxSYJ
11:05:26 sigq: Title: Private Paste :: Paste cegAPPxSYJ ( at pgsql.privatepaste.com )
11:05:29 pimpaa: the first if works just fine
11:05:38 pimpaa: the two inner ifs dont
11:05:50 pimpaa: they dont return me anything ever
11:09:57 pimpaa: actually, testing this out
11:10:04 pimpaa: just the "cruzamento" if doest not work
11:36:20 darkblue_B: alert! alert!
11:36:36 darkblue_B: Postgres Day at OSCon, speaking slot on PostGIS
11:36:48 darkblue_B: I am committing to do the talk
11:37:03 darkblue_B: I have to name it and provide an outline
11:37:38 darkblue_B: If anyone has a burning desire to see some particular topic hilited in front of a bunch of Postgres developers please speak up
11:38:17 darkblue_B: I am leaning towards "What works with Postgres - the Open Geo Data interoperabilty overview"
11:38:26 darkblue_B: something like that
12:36:16 darkblue_B: hi pramsey
12:36:43 darkblue_B: two things quickly
12:37:16 darkblue_B: one is that I have been offered a speaking slot at the Postgres Day at OReilly OSCon in July
12:37:32 darkblue_B: I have to give them a talk title and small outline today to get it though
12:37:46 darkblue_B: PostGIS related, of course
12:38:15 darkblue_B: I am leaning towards "What works with Postgres - the Open Geo Data interoperabilty overview"
12:39:07 darkblue_B: the organizer said that they wanted to get away from Intro or Overview sort of things
12:39:26 pramsey: OK
12:41:31 darkblue_B: if you had some other thought on what needs to be hilited in front of a bunch of Postgres core developers, please say so
12:41:54 darkblue_B: its not all committers, but a lot of committers tend to show up to that I think
12:42:05 darkblue_B: -
12:42:25 darkblue_B: the second thing was I read your recent post on MySQL and Pg benchmark
12:43:02 darkblue_B: what I take away from that is that a correctly configured MySQL is now very compeitive with a PostGIS setup for read only, basic spatial queries
12:43:09 darkblue_B: and basic r/w too
12:43:15 darkblue_B: this is pretty surprising
12:43:34 darkblue_B: and not a small matter.. zillions of people work with MySQL as a matter of course for speed alreday
12:44:09 darkblue_B: now that spatial is hot.. it just means that Pg doesnt hold a privelaged position in this space, all of a suddent
12:46:25 darkblue_B: Pg by any objective measure holds a unique place for huge, truly huge, data stores
12:46:37 darkblue_B: but how many people do that.. limited
12:46:55 darkblue_B: so it just seems like a big developement
13:31:10 pramsey: darkblue_B: hm, perhaps I was too kind in my wording.
13:31:29 pramsey: Basically the state-of-the-art in MySQL spatial hasn't changed in 5 years.
13:31:50 pramsey: It still does sweet f. all in terms of features.
13:32:01 pramsey: And in terms of throughput, PostGIS is 50% faster.
13:32:18 pramsey: And that's for very simple things.
13:32:45 pramsey: For even a minor spatial database query ( spatial join ) it falls to bits ( http://blog.cleverelephant.ca/2009/06/mysql-snark-2.html )
13:32:47 sigq: Title: Paul Ramsey: MySQL Snark #2 ( at blog.cleverelephant.ca )
13:33:58 pramsey: so, darkblue_B, your analysis is correct, but it's correct as of 5 years ago, when MySQL first added spatial. That addition removed a unique feature, to the extent that the MySQL users could fit their use cases into the narrow functionality box provided.
13:34:50 pramsey: And many who tried to do so later found they had made a bad choice http://blog.redfin.com/devblog/2007/11/elephant_versus_dolphin_which_is_faster_which_is_smarter.html
13:34:51 sigq: Title: Elephant versus Dolphin: Which is Faster? Which is Smarter? | Redfin Developers Blog ( at blog.redfin.com )
13:37:28 darkblue_B: hmmm 50% faster .. I didnt get that from the graphs
13:37:34 darkblue_B: I will look again
13:37:54 pramsey: well, at 18 threads : ) I'm cherry picking : )
13:38:03 darkblue_B: yeah
13:38:19 pramsey: anyways, it's better that MySQL on MySQL's strongest possible turf
13:38:25 pramsey: "we're fast!"
13:38:36 darkblue_B: right
13:38:43 darkblue_B: ok, thats helpful
13:38:53 pramsey: on all the rest of the turf ( are you reliable, can you handle complex stuff, are you ACID ) it's no contest
13:39:00 darkblue_B: yes
13:39:57 darkblue_B: the OSGeo booth will be right next to eh Drizzle booth at OSCon
13:40:13 : * pramsey weeps to miss yet another OSCon.
13:40:26 darkblue_B: so I will get a chance to ask someone knowledgeable about which fork is going to carry the Geo torch
13:40:33 darkblue_B: OSCon is an odd show
13:40:41 darkblue_B: the quality of the presentations is very high..
13:40:48 darkblue_B: its almost akin to broadcast television
13:40:55 darkblue_B: TED.. that sort of thing
13:41:19 darkblue_B: this is the first time it will be in San Jose, so no one knows what the attendance will be like.. high or low
13:41:31 pramsey: Oh, really, I didn't realize they had left Portland.
13:41:50 darkblue_B: but in the heart of Silicon Valley, with budget pressure favoring OSS, and against travel budgets.. its a toss up
13:41:54 darkblue_B: yes, San Jose
13:42:02 darkblue_B: part of why I am eager to jump in..
13:42:06 darkblue_B: no hotel
13:42:48 darkblue_B: now admittedly, the slot I am getting is on PgDay, not on the main schedule
13:43:01 darkblue_B: I think they have 2 Pg talks on the main rotation
13:43:08 darkblue_B: thats 100 people or more without question
13:43:18 darkblue_B: this Pg day.. it'll be the Pg faithful
13:43:32 darkblue_B: as I say, probably more than a few committers
13:43:54 darkblue_B: but no idea on total head count
13:44:16 darkblue_B: as far as a talk, I have until midnight to get them an outline
13:44:21 pramsey: : )
13:44:25 darkblue_B: I think I'll stick to the stack idea
13:44:46 darkblue_B: I *would* like to say something about these MySQL benches
13:44:56 darkblue_B: I appreciate your mitigating view there
13:45:05 darkblue_B: its not what I took away at first glance
13:45:24 darkblue_B: and.. a chance to wave the OSGeo flag a bit
13:45:31 darkblue_B: I think thats good..
13:45:43 darkblue_B: not just boosterism, but actualy substantially good
13:45:48 darkblue_B: the stack make apps
13:47:17 pramsey: and apps are what people use
13:47:30 darkblue_B: exacta-mundo
13:55:20 darkblue_B: oh - that redfin post is the best I've seen
13:55:27 darkblue_B: thats very informative
13:56:04 darkblue_B: in my 'research' I havent even made mixed spatial/non-spatial indexes, let alone test them
13:56:34 darkblue_B: this is substance for a whole section of the talk...
18:58:34 ltcabral: any fast way to order file lines lexicographically? : )
19:00:04 ltcabral: oops wrong channel sry