#POSTGIS IRC Log - 2008-06-05

For logs after Feb 3, 2007, all times are GMT-8. Prior logs are GMT-9.
Back to Logs
03:21:41 CIA-6: mcayland * r2799 /trunk/lwgeom/ ( 10 files ): Blast away the huge amounts of legacy C code related to PostgreSQL versions < 8.1. Next step will be to work on the SQL script files too.
03:53:55 CIA-6: mcayland * r2800 /trunk/lwgeom/lwpostgis.sql.in.c: Remove all of the pre-PostgreSQL 7.3 schema #ifdefs from lwpostgis.sql.in to make things readable once again.
04:10:19 CIA-6: mcayland * r2801 /trunk/loader/ ( Makefile.pgsql2shp.in PQunescapeBytea.c ): Remove PQunescapeBytea hack from pgsql2shp which is no longer needed, as it was only required for versions of PostgreSQL < 7.3
04:22:11 CIA-6: mcayland * r2802 /trunk/lwgeom/ ( long_xact.sql.in sqldefines.h.in ): Remove the HAS_SCHEMA #define and all related #else code since we now guarantee to be using PostgreSQL > 7.2
05:12:34 CIA-6: mcayland * r2803 /trunk/lwgeom/ ( long_xact.sql.in lwpostgis.sql.in.c sqlmm.sql.in ): Switch all SQL and PL/PGSQL stored procedures over to use dollar quoting instead of apostrophes for function bodies. At last, most of the stored procedures actually become readable
07:08:51 jalmeida: is possible works to postgis and javascript
07:09:28 jalmeida: renderer the features on browser( client ) from database
07:09:42 jalmeida: ??
07:33:33 CIA-6: mcayland * r2804 /trunk/ ( configure.ac configure.in macros/ac_proj4_version.m4 ): Update copyrights for configure.ac ( rewritten from scratch, and now renamed from the deprecated form configure.in ) and also the PROJ.4 version detection macro
07:43:57 bitner: jalmeida: you cannot work directly with the database from postgis, you must have some server side code sitting between the browser and the database like http://featureserver.org or http://geoserver.org
07:43:58 sigq: Title: FeatureServer -- RESTful Geographic Feature Storage ( at featureserver.org )
07:44:39 bitner: if all you want to do is render, you could also render as an image using mapserver etc
07:50:10 CIA-6: mcayland * r2805 /trunk/ ( 6 files in 2 dirs ): Update new Makefile copyrights, also remove compat.h from the loader/ directory since it is now no longer needed.
07:51:41 jalmeida: do you have to example to work with openlayers and postgis ????
07:52:07 CIA-6: mcayland * r2806 /trunk/lwgeom/Makefile: Remove Makefile from lwgeom/ directory; it is no longer needed as it is automatically generated from lwgeom/Makefile.in
07:52:15 bitner: featureserver and mapserver work very well between openlayers and postgis
07:52:21 jalmeida: i create my database and want editing my featurers using openlayers, its is possible ???
07:52:33 bitner: with featureserver, yes
07:52:34 jalmeida: bitner, = )
09:11:02 CIA-6: mcayland * r2807 /trunk/lwgeom/lwpostgis.sql.in.c:
09:11:02 CIA-6: Some more updates to lwpostgis.sql.in.c; correct a couple of missing conversions
09:11:02 CIA-6: from OPAQUE_TYPE to internal ( how on earth did regression tests pass with this
09:11:02 CIA-6: still in place?? ), plus remove the UPDATE on pg_opclass to change the OID of the
09:11:02 CIA-6: index storage type since we can do this by specifying a STORAGE clause instead.
09:21:31 pagameba: hi - looking for ways to optimize my postgis experience with mapserver
09:21:33 pagameba: I have the standard gist index on the_geom
09:21:35 pagameba: my layers are classified on a field in the data - fcode
09:21:39 pagameba: I have an index on fcode
09:21:43 pagameba: its still somewhat slow
09:22:41 pagameba: should I use a subselect in the DATA statement to limit the query ie the_geom from ( select * from mytable where fcode in ( 'val1', 'val2', 'val3' ) ) as foo using unique gid using srid=-1
09:23:09 bitner: pagameba: best way is to intercept the query and run it using explain analyze with psql or something like pgadmin
09:23:37 bitner: I typically will do something like that
09:24:01 bitner: using subselects that is
09:24:17 pagameba: bitner thx
09:24:39 pagameba: is it useful to explicitly set the fields to return as well instead of using *?
09:25:09 bitner: I don't think that should matter unless there are a ton of fields
09:25:27 pagameba: k
09:26:58 bitner: the trick I use with mapserver is to use a pl/pgsql function that takes in data in %% notation in my mapfile and then build the correct query in pl/pgsql so that I can easily have optional fields and the like
09:27:52 : * pagameba futzes with shp2img
09:27:59 bitner: ie DATA the_geom from ( select * from myfunction( '%val1%','%val2%' ) )using ...
09:47:00 pagameba: it does seem to improve things by using the subselect
09:47:36 bitner: did you try explain analyze?
09:48:57 bitner: that will tell you for sure how you are using your indexes
09:50:07 pagameba: what is a bitmap heap scan?
09:51:34 bitner: a lot of the bitmap useage came in to play at 8.2
09:51:49 bitner: main thing they allow is using multicolumn indexes out of order
09:52:09 bitner: or making use of multiple separate indexes from one query
09:52:13 pagameba: I only see it using one index
09:53:52 bitner: are you just running the subquery or are you running the actual query that mapserver is calling?
09:54:01 pagameba: both
09:54:08 pagameba: the subquery doesn't seem to use any index
09:54:21 pagameba: it is just doing a seq scan
09:54:25 pagameba: with a filter
09:54:53 bitner: what if you just explain the subquery? does that use the index on fcode?
09:55:05 pagameba: no
09:55:11 pagameba: select the_geom, fcode, gid from dtb_hydro_polygon where fcode in ( 'WASW','WARVLK','WARVDL','WALK' )
09:55:29 pagameba: does a seq scan
09:55:57 pagameba: I have a btree index on fcode
09:56:05 bitner: you have index on fcode right? have you vacuum analyzed the index?
09:57:09 : * pagameba does that
09:57:26 bitner: in psql or pgadmin query window or whatever run set enable_seq_scan to off; select the_geom......
09:57:48 bitner: to try to force the query planner to try to use index
09:58:08 bitner: does fcode='WASW' rather than the in statement use index?
09:58:39 pagameba: doesn't seem to change it
09:59:07 pagameba: set enable_seq_scan to off
09:59:17 pagameba: yields unrecognized configuration parameter
09:59:53 bitner: drats, memory must be faulty
10:00:04 : * bitner looks up what the actual statement is
10:01:23 bitner: set enable_seqscan to off
10:01:25 bitner: oops
10:02:18 pagameba: ok, now it uses the index
10:02:47 bitner: for some reason your planner does not think it is worthwhile to use the index
10:03:12 bitner: is it faster when it uses the index?
10:04:12 pagameba: difficult to say, but I think so ( marginally )
10:04:52 : * pagameba steps out for a meeting
10:04:54 pagameba: thx bitner
10:05:00 bitner: np, cheers
11:36:07 cgs_bob: hello all. I am trying to find a way to display borehole data and it seems that I can use ST_MakeBox3D. Unfortunately the examples I've found ( such as ST_SetSRID( ST_MakeBox3D( ST_MakePoint( west_in,south_in ), ST_MakePoint( east_in,north_in ) ),4326 ) ) do not show a box with depth. anyone have any pointers?
11:44:14 bitner: "do not show a box with depth" ??
11:44:30 bitner: you are not feeding makepoint any z values
11:51:10 cgs_bob: bitner: I was looking at the docs to postgis and did not understand what LLB and URT in ST_MakeBox3D( <LLB>, <URT> ) is. I was just wishing to find examples on the net so I could figure out those arguments.
11:51:39 bitner: Lower Left Bottom, Upper Right Top
11:53:22 bitner: do you have your boring data in the db as points?
11:54:15 cgs_bob: thanks. now I'd like to make a suggestion for the docs....a table of abbreviations and glossary : )
11:56:55 cgs_bob: bitner: I have 25000 boreholes in a database with ground water depth, lithology, sample analysis, and various other geotechnical data. btw, a lot of this data is also available over the internet
12:02:00 cgs_bob: location of the borehole is in sample_location's top_lat,top_lon and the lithology has top_depth, bottom_depth for each layer
21:17:07 chris123: greets
21:18:17 chris123: currently setting up postgis with postgresql and I keep getting the same error when I enter the following command
21:18:26 chris123: psql -d gis -f /usr/share/postgresql-8.3-postgis/lwpostgis.sql
21:19:06 chris123: the error is as follows: psql:/usr/share/postgresql-8.3-postgis/lwpostgis.sql:1557: ERROR:
21:19:20 chris123: any insight on how I can track this one down??
21:22:21 chris123: tks its late so need some sleep, will try again tomorrow
21:22:22 chris123: ciao
21:34:22 simplexio: chris123: did you do createlang plpgsql ?
21:59:46 chris123: <simplexio> yes
22:26:21 cgs_bob: hello all. I'm looking for a program that would allow you to create a postgis query and display the results without having to create new tables. is there such a beast?
22:49:24 springmeyer: cgs_bob: what do you mean? are you trying to continuously write over a table?
22:52:38 cgs_bob: springmeyer: I'm just looking for a program that would make it easy for me to learn postgis. for instance, in qgis, I can create a query for a certain feature, but you can't do joins.
22:53:49 springmeyer: ah, i see... ya, so you are not writing queries 'by hand' in pure sql, you are interacting with postgis through qgis...
22:53:53 springmeyer: hmm...
22:54:23 springmeyer: well, particularly to joins in qgis... I think there is a plugin now that will do them
22:54:39 : * springmeyer digs around to verify
22:55:36 cgs_bob: cool...I'll look for it too
22:57:24 nhv: wow can't you just make layers from queries ??
22:57:36 springmeyer: http://www.geog.uvic.ca/spar/carson/
22:57:37 sigq: Title: SPAR! Spatial Pattern Analysis & Research Lab -- Carson Farmer ( at www.geog.uvic.ca )
22:58:41 springmeyer: in terms of learning postgis... cgs_bob you'll really benefit from spending some time getting to know how to write basic select queries using the psql interpreter
23:01:49 cgs_bob: springmeyer: I have been working on a query that would allow me to look at my borehole database, but the was getting weird results. so, I thought I'd look for tools that would make it easy to play with postgis
23:02:10 springmeyer: ah, sounds good
23:13:35 springmeyer: cgs_bob: whats your data format in general?
23:18:11 cgs_bob: springmeyer: I have a table that has the location of each borehole along with other info about that borehole. I also have a lithology table that has a boreholeid, top_depth,bottom_depth,lith_type, and strat_unit_name. I'm trying to find a way to visualize this.
23:19:37 springmeyer: so, you've got latitude / longitude in your table... or you've imported into postgis and the data is stored in the standard wkb?
23:21:49 springmeyer: and what format do you store it in outside of postgis?
23:23:16 springmeyer: seems like if your boreholes are 3d then you're on the right track storing your data in postgis since postgis supports Z values, but you'll need to serve your data out to a 3D client
23:23:31 springmeyer: why not Google Earth?
23:24:57 cgs_bob: the first table I mentioned has a geometry table and the lithology table has z values associated with the different layers of the earth. I'm hoping to make a linestring feature that I will eventually put into a shapefile
23:27:01 cgs_bob: I work for a state agency and I have been able to convince my boss to take a look at FOSS. he wants to see what I can do first
23:28:00 cgs_bob: I can't load software on my work computer, so I have to give him the results in a form that can be imported into our gis
23:28:58 nhv: what is your gis ??
23:30:04 nhv: or what formats does it ingest ?
23:32:15 : * nhv 's first job as a 'coder' was visualizing the boreholes in Alberta
23:32:49 springmeyer: nhv sounds like your man
23:33:30 : * nhv is not a QGis pro though
23:33:36 : * springmeyer having not worked with boreholes wonders what cgs_bob wants to visualize...
23:33:48 : * nhv too
23:34:00 : * springmeyer is headed out for the night however
23:34:21 cgs_bob: we have Intergraph MGE and have plans to migrate to Geomedia Professional. Unfortunately MGE uses a proprietary file format but Geomedia can import shapefiles
23:34:45 nhv: can you export shapefiles ?
23:36:11 nhv: or better can you make a 'SQL View' of your data
23:37:37 cgs_bob: one other thing I have to say. We use to use a software package called ERMA Site Geologist that allows you to visualize boreholes. we can't use it any more because our gis is obsolete, so we are looking for other options.
23:40:02 cgs_bob: I have all of the tables in postgresql now.
23:40:22 cgs_bob: anyways, I got to get going...toddler woke up : )