Slow spatial query in oracle 11.2.0.3 [WAS 10.2.0.4 to 11.2.0.1 upgrade = SLOW sdo_geometry return]
Hi,
I've found a discussion ended last year.
I've Oracle 11.2.0.3 with the same problem: some query are very slow (in the Excecution plan I've found "PX COORDINATOR")
Others installation of Oracle 11.2.0.3 haven't same problem: in the excecution plan found "DOMAIN INDEX".
I 've solved using statement "ALTER TABLE table_name NOPARALLEL;" which disable parallel feature in the table
Can anyone tell me if it's a good (bad) solution ? thank you
Bye
Gabriele
Gabriele,
As indicated in the original post - the following work-around might work for you for bug 9743250. What this does is essentially force the optimizer to ALWAYS use the spatial index. That is usually good - unless the query is actually more efficient not using the index, such as when it is much cheaper to just to a FTS. Anyhow this is what you can do from 11.2.0.1 to 11.2.0.3 - fixed in 11.2.0.4 and 12.1.0.1:
connect /as sysdba
alter session set current_schema=MDSYS;
DISASSOCIATE STATISTICS FROM INDEXTYPES spatial_index FORCE
DISASSOCIATE STATISTICS FROM PACKAGES sdo_3gl FORCE;
DISASSOCIATE STATISTICS FROM PACKAGES prvt_idx FORCE;
Bryan
Similar Messages
-
We are finally getting ready to move our applications from 10gR2 to 11gR2, and first off, I ran into a complete show stopper. This is with an in-place upgrade on Solaris x86-64.
On a query using SDO_FILTER, we can:
1. quickly count the number of records that would be returned
2. quickly return all column data from the table EXCEPT sdo_geometry data
Without the SDO_FILTER (using just ROWNUM for example), sdo_geometry data is returned quickly.
I've dug around google and metalink without hitting on any suggestions. The table data we are querying is partitioned into 9 groups by range (PK). The spatial index is not partitioned.
Here is a sample query:
SELECT COMPATH_ID PK,
GEOMETRY_L GEOM,
PATH_TYP_D,
CABINS_D,
COVERDEPTH,
DEPTH_U_D,
DATE_INST,
DISPOSTN_D,
MAPSRC_D,
MAPACU,
RECUPDATE,
GEOLOC,
UPDTGEOLOC,
UPDTBY,
GEOM_LENGTH,
PATH_CNT_D,
H_LEVEL,
QCD_DATE,
QCD_BY
FROM CVC_ASIS.C_PATH_SEGMENT A
WHERE SDO_FILTER (A.GEOMETRY_L,
SDO_GEOMETRY (2003,
8307,
NULL,
SDO_ELEM_INFO_ARRAY (1, 1003, 3),
SDO_ORDINATE_ARRAY (-97.3618,
35.398,
-97.352,
35.41))) =
'TRUE'
AND ROWNUM < 200;(There are 199 items - I just added the rownum to keep it from being a runaway query)
Without the SDO_GEOMETRY column geometry_l returned, I get this plan - FAST:
199 rows selected.
Elapsed: 00:00:01.82
Execution Plan
Plan hash value: 1088531910
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | Pstart| Pstop |
| 0 | SELECT STATEMENT | | 199 | 29054 | 0 (0)| 00:00:01 | | |
|* 1 | COUNT STOPKEY | | | | | | | |
| 2 | TABLE ACCESS BY GLOBAL INDEX ROWID| C_PATH_SEGMENT | 3804 | 542K| 0 (0)| 00:00:01
|* 3 | DOMAIN INDEX | C_PATH_SEGMENT_SNDXL | | | 0 (0)| 00:00:01 | | |
Predicate Information (identified by operation id):
1 - filter(ROWNUM<200)
3 - access("MDSYS"."SDO_FILTER"("A"."GEOMETRY_L","MDSYS"."SDO_GEOMETRY"(2003,8307,NULL,"SDO_ELEM_
03,3),"SDO_ORDINATE_ARRAY"((-97.3618),35.398,(-97.352),35.41)))='TRUE')
Statistics
46 recursive calls
2 db block gets
254 consistent gets
0 physical reads
0 redo size
22543 bytes sent via SQL*Net to client
349 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
2 sorts (memory)
0 sorts (disk)
199 rows processedWith the SDO_GEOMETRY column geometry_l returned, I get this plan (ouch!):
199 rows selected.
Elapsed: 00:00:51.44
Execution Plan
Plan hash value: 2419072205
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | Pstart| Pstop | TQ |IN-OUT
| 0 | SELECT STATEMENT | | 199 | 29054 | 356 (1)| 00:00:05 | | | | | |
|* 1 | COUNT STOPKEY | | | | | | | | | | |
| 2 | PX COORDINATOR | | | | | | | | | | |
| 3 | PX SEND QC (RANDOM) | :TQ10000 | 3804 | 542K| 356 (1)| 00:00:05 | | | Q1,00 | P->
|* 4 | COUNT STOPKEY | | | | | | | | Q1,00 | PCWC | |
| 5 | PX BLOCK ITERATOR | | 3804 | 542K| 356 (1)| 00:00:05 | 1 | 9 | Q1,00 |
|* 6 | TABLE ACCESS FULL| C_PATH_SEGMENT | 3804 | 542K| 356 (1)| 00:00:05 | 1 |
Predicate Information (identified by operation id):
1 - filter(ROWNUM<200)
4 - filter(ROWNUM<200)
6 - filter("MDSYS"."SDO_FILTER"("A"."GEOMETRY_L","MDSYS"."SDO_GEOMETRY"(2003,8307,NULL,"SDO_ELEM_
_ORDINATE_ARRAY"((-97.3618),35.398,(-97.352),35.41)))='TRUE')
Statistics
227579 recursive calls
0 db block gets
8764708 consistent gets
0 physical reads
0 redo size
85130 bytes sent via SQL*Net to client
2455 bytes received via SQL*Net from client
29 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
199 rows processedIdeas?François.
I would think there would be a significant number of licensed users for spatial. Of course there is little in the way to count all the locator users, who probably number 1000x more. Oracle's internal BI products use it for their map components for goodness sakes, and should be equally effected by this issue. I can't believe they would apply a hint for every single spatial select this effects.
Why in the world this issue was not made priority #1 by the product team and corrected quickly is beyond belief. Unfortunately, my thought is that since Oracle now has so many pies in the oven, the database and the specialty parts (like spatial) are on the back burner. And I would guess that the product managers get rewarded more for shiny new features, than keeping existing features working correctly.
While they claim it is corrected in 12, that does all of us in the real world using actual shipping products no good whatsoever. Oracle hint: This is a good way to lose customers.
After experiencing a few 10.2.0.4 issues from back-ported code from 11g that changed the expected behavior of the results (with no documentation indicating such changes), we created a test suite to test all database patches for broken/changed spatial issues, and found several that we had to work around. I must say I am no longer comfortable about anything coming from Oracle working as advertised, sad to say. I'm now completely in the Show-Me camp for any feature, new or old.
My professional suggestion is to continue to use 10gR2 (which is based on mostly solid code written by people I knew) if you can, and ignore 11g altogether because of this long overdue issue. When 12 comes out, hopefully Oracle will gain back some respect from me, but I'm not real optimistic on that point.
Bryan
Spatial Architect -
Oracle 9i 10g slow spatial query
I have been using the following spatial query in oracle 9i to assign jurisdictions to spatial points.
SELECT /*+ordered*/ L.TCH_X_COORD, L.TCH_Y_COORD, T.NAME
FROM tis.tis_coordinates_hold_on L,
geospatial.mdt_admin_districts T
WHERE SDO_RELATE(T.GEOMETRY,
MDSYS.SDO_GEOMETRY(2001,41079,MDSYS.SDO_POINT_TYPE(tch_x_coord,tch_y_coord,NULL),
NULL,NULL),'mask=ANYINTERACT querytype=WINDOW') = 'TRUE'
AND tch_assigned_base_rte='M00001N'
By using the ordered hint I was able to double the retrieval rate. My problem is this: We are now migrating to oracle 10g, and the use of the ordered returns no rows (different incorrect results). If I remove the + ordered the results are correct, but I no longer get the increased speed of retrieval. Has anyone run into this problem, and can you suggest a way that I could speed up this query. As it stands now it takes more than 1/2 sec for each record retrieved, and I have millions of records !! Thanks.We're using 10.2.4.0 I believe I have found the reason the queries were different. When I dropped the spatial index and re-created it, the query returned the same rows as 9i. However, I still have an issue with the speed. 10g appears to be about 1/3 as fast as 9i was. It doesn't seem to make any difference whether I put in the +ordered or not. Our dba says just based on cpu and processing speed the server that has 10g should be significantly faster. Any comments? Thanks.
-
Poor performance with Oracle Spatial when spatial query invoked remotely
Is anyone aware of any problems with Oracle Spatial (10.2.0.4 with patches 6989483 and 7003151 on Red Hat Linux 4) which might explain why a spatial query (SDO_WITHIN_DISTANCE) would perform 20 times worse when it was invoked remotely from another computer (using SQLplus) vs. invoking the very same query from the database server itself (also using SQLplus)?
Does Oracle Spatial have any known problems with servers which use SAN disk storage? That is the primary difference between a server in which I see this poor performance and another server where the performance is fine.
Thank you in advance for any thoughts you might share.OK, that's clearer.
Are you sure it is the SQL inside the procedure that is causing the problem? To check, try extracting the SQL from inside the procedure and run it in SQLPLUS with
set autotrace on
set timing on
SELECT ....If the plans and performance are the same then it may be something inside the procedure itself.
Have you profiled the procedure? Here is an example of how to do it:
Prompt Firstly, create PL/SQL profiler table
@$ORACLE_HOME/rdbms/admin/proftab.sql
Prompt Secondly, use the profiler to gather stats on execution characteristics
DECLARE
l_run_num PLS_INTEGER := 1;
l_max_num PLS_INTEGER := 1;
v_geom mdsys.sdo_geometry := mdsys.sdo_geometry(2002,null,null,sdo_elem_info_array(1,2,1),sdo_ordinate_array(0,0,45,45,90,0,135,45,180,0,180,-45,45,-45,0,0));
BEGIN
dbms_output.put_line('Start Profiler Result = ' || DBMS_PROFILER.START_PROFILER(run_comment => 'PARALLEL PROFILE')); -- The comment name can be anything: here it is related to the Parallel procedure I am testing.
v_geom := Parallel(v_geom,10,0.05,1); -- Put your procedure call here
dbms_output.put_line('Stop Profiler Result = ' || DBMS_PROFILER.STOP_PROFILER );
END;
SHOW ERRORS
Prompt Finally, report activity
COLUMN runid FORMAT 99999
COLUMN run_comment FORMAT A40
SELECT runid || ',' || run_date || ',' || run_comment || ',' || run_total_time
FROM plsql_profiler_runs
ORDER BY runid;
COLUMN runid FORMAT 99999
COLUMN unit_number FORMAT 99999
COLUMN unit_type FORMAT A20
COLUMN unit_owner FORMAT A20
COLUMN text FORMAT A100
compute sum label 'Total_Time' of total_time on runid
break on runid skip 1
set linesize 200
SELECT u.runid || ',' ||
u.unit_name,
d.line#,
d.total_occur,
d.total_time,
text
FROM plsql_profiler_units u
JOIN plsql_profiler_data d ON u.runid = d.runid
AND
u.unit_number = d.unit_number
JOIN all_source als ON ( als.owner = 'CODESYS'
AND als.type = u.unit_type
AND als.name = u.unit_name
AND als.line = d.line# )
WHERE u.runid = (SELECT max(runid) FROM plsql_profiler_runs)
ORDER BY d.total_time desc;Run the profiler in both environments and see if you can see where the slowdown exists.
regards
Simon -
Hello, I have this problem,
I use php for read data from oracle table,
all works right, I have any problem whith varchar number type data, but when I must read geometry data like
(type MDSYS.SDO_GEOMETRY)
I can't display it on the web page.
esemple of spatial query that I use:
select GEOMETRY_1
from table_a
where
ID = 970;
If I use sql plus I have this result
SDO_GEOMETRY(3001, NULL, NULL, SDO_ELEM_INFO_ARRAY(1, 0, 6000, 4, 1, 1), SDO_ORDINATE_ARRAY(1, -0, -
0, 580094, 4998494, -1))
but blank page when I use php sentence.
I use PHP Version 4.4.0
modules load are:
extension=php_sdo.dll
extension=php_oci8.dll
But I don't know which other module I need and also which php function to use for read array of array. at the moment I use @OCIFetchInto.
Someone can help me?
My English is very bad please reply easy.
Thanks for all!
AngeloThank you very much CJ.
Now there isn't any problems: you and this link (Re: PHP and spatial data and Adamo Bozzetti helped me
I have used SDO_UTIL.GETVERTICES
my complete php script is:
$conn = @OCILogon("xxxxxx", "xxxxx", "xxxxx") or die
( "Non riesco a connettermi al server $host ");
$query = "select
TABLE_A.tipo_via,
TABLE_A.nome_via,
c.codice_via,
t.X,
t.Y
from TABLE B c, TABLE_A , TABLE(SDO_UTIL.GETVERTICES(c.geometry1)) t
where c.codice_via = 830
and c.codice_via = TABLE_A.id_via
$istruzione = @OCIParse($conn, $query);
@OCIExecute($istruzione);
$nrows = @OCIFetchInto($istruzione, $results);
$trovati = 0;
while(@OCIFetchInto($istruzione, $results))
$trovati++;
?>
<table border width="800" >
<tr>
<td align="center" width="80" bgcolor ="#aaffee"><?echo trim($results[0]);?></td>
<td align="center" width="150" bgcolor ="#eeffee"><?echo trim($results[1]);?></td>
<td align="center" width="50" bgcolor ="#eeffee"><?echo trim($results[2]);?></td>
<td align="center" width="50" bgcolor ="#eeffee"><?echo trim($results[3]);?></td>
<td align="center" width="50" bgcolor ="#eeffee"><?echo trim($results[4]);?></td>
</tr>
</table>
<?
} //fine While
@OCIFreeStatement($istruzione);
@OCILogOff($conn); -
Oracle spatial query SDO_LRS with php
Hello, I have this problem,
I use php for read data from oracle table.
I know I can't show on web page the result of SDO_ORDINATE_ARRAY
Christopher Jones has written that I must use SDO_UTIL.GETVERTICES like as
select
t.X,
t.Y
from table_a, TABLE(SDO_UTIL.GETVERTICES(table_a.GEOMETRY_1)) t
where
id_num = 970;
But now I must read a projection point ordinate of a point with respect to a geometric segment
and the same problem come back.
esemple of spatial query that I use:
SELECT
SDO_LRS.PROJECT_PT(geometry,
SDO_GEOMETRY(3301, NULL, NULL, SDO_ELEM_INFO_ARRAY(1, 1, 1), SDO_ORDINATE_ARRAY(580094.69, 4998493.68, NULL)) )
FROM
TABLE_B
WHERE
ID_CODE = 970;
If I use sql plus I have this result
SDO_LRS.PROJECT_PT(GEOMETRY,SD(SDO_GTYPE, SDO_SRID, SDO_POINT(X, Y, Z), SDO_ELEM
SDO_GEOMETRY(3301, NULL, NULL, SDO_ELEM_INFO_ARRAY(1, 1, 1), SDO_ORDINATE_ARRAY(
580093,325, 4998497,57, 45,2254405))
Which is exacly the query for extract these coordinates?
Someone can help me?
My English is very bad please reply easy.
Thanks for all!
AngeloHello thanks for all to everyboby.
I put the same question on Forum /Category/Database/Spatial and I have received this reply
from Brian Camire, this query worked well.
SELECT
T.X,
T.Y
FROM
TABLE
SELECT
SDO_UTIL.GETVERTICES
SDO_LRS.PROJECT_PT
GEOMETRY,
SDO_GEOMETRY
3301,
NULL,
NULL,
SDO_ELEM_INFO_ARRAY(1, 1, 1),
SDO_ORDINATE_ARRAY(580093.859, 4998493.5, NULL)
FROM
TABLE_B
WHERE
ID_CODE = 970
) T;
Best regards
Angelo. -
Spatial query w/ logical operators in a subquery on Linux using Oracle 1
Hello...
I'm being given XML that I need to parse and create SQL from it. Below my signature is the SQL that I generate (its a simple example) and can be more elaborate due to the logical operators (AND and OR) given in the XML.
The problem that we're seeing is performance-based: when one of the logical operator queries is Spatial-based, like below. Running it in JDBC, the query never returns.
The spatial query just by itself returns ~300 rows (takes approximately 0.2 seconds). The "PERCENTAGE" query by itself returns about ~40000 rows (takes approximately ~.8 seconds).
If we run the query below with 2 non-spatial subqueries, the result returns and performance is very acceptable (~ 0.9 seconds)-- the result set is about 80000 rows.
Thanks,
Jim
=========================
SELECT
COLUMN_WE_WANT , RESULTS
FROM
TABLE_A
WHERE
COLUMN_WE_WANT IN
SELECT
COLUMN_WE_WANT
FROM
TABLE_A
WHERE
SDO_OVERLAPBDYINTERSECT(TABLE_A.MY_SPATIAL_COLUMN,
SDO_GEOMETRY(2003,
4326,
null,
SDO_elem_info_array( 1 , 3 , 1 ),
SDO_ORDINATE_ARRAY( lng_1,lat_1 , lng_2,lat_2 , lng_3,lat_3 , lng_4,lat_4 , lng_1,lat_1 )
) = 'TRUE'
OR
COLUMN_WE_WANT IN
SELECT
COLUMN_WE_WANT
FROM
TABLE_B
WHERE
SOME_PERCENTAGE_RATE_COLUMN < 90
)Its difficult to comment without seeing the execution plan. You should trace this query to get a better idea of exactly what's happening.
Depending on the complexity of the logical operators, I would look at doing this using set operations. So for an OR you might have...
SELECT COLUMN_WE_WANT, RESULTS
FROM TABLE_A
WHERE COLUMN_WE_WANT IN (
SELECT COLUMN_WE_WANT
FROM TABLE_A
WHERE SDO_OVERLAPBDYINTERSECT(TABLE_A.MY_SPATIAL_COLUMN,SDO_GEOMETRY(2003,4326,
NULL, SDO_elem_info_array( 1 , 3 , 1 ),
SDO_ORDINATE_ARRAY( lng_1,lat_1 , lng_2,lat_2 , lng_3,lat_3 , lng_4,lat_4 , lng_1,lat_1 ) )) = 'TRUE'
UNION
SELECT COLUMN_WE_WANT
FROM TABLE_B
WHERE SOME_PERCENTAGE_RATE_COLUMN < 90)For an AND, you would use INTERSECT.
Edited by: Reggie to remove the extra INTERSECT -
Spatial query w/ logical operators in a subquery on Linux using Oracle 10g
Hello...
I'm being given XML that I need to parse and create SQL from it. Below my signature is the SQL that I generate (its a simple example) and can be more elaborate due to the logical operators (AND and OR) given in the XML.
The problem that we're seeing is performance-based: when one of the logical operator queries is Spatial-based, like below. Running it in JDBC, the query never returns.
The spatial query just by itself returns ~300 rows (takes approximately 0.2 seconds). The "PERCENTAGE" query by itself returns about ~40000 rows (takes approximately ~.8 seconds).
If we run the query below with 2 non-spatial subqueries, the result returns and performance is very acceptable (~ 0.9 seconds)-- the result set is about 80000 rows.
Thanks,
Jim
=========================
SELECT
COLUMN_WE_WANT , RESULTS
FROM
TABLE_A
WHERE
COLUMN_WE_WANT IN
SELECT
COLUMN_WE_WANT
FROM
TABLE_A
WHERE
SDO_OVERLAPBDYINTERSECT(TABLE_A.MY_SPATIAL_COLUMN,
SDO_GEOMETRY(2003,
4326,
null,
SDO_elem_info_array( 1 , 3 , 1 ),
SDO_ORDINATE_ARRAY( lng_1,lat_1 , lng_2,lat_2 , lng_3,lat_3 , lng_4,lat_4 , lng_1,lat_1 ) )
) = 'TRUE'
OR
COLUMN_WE_WANT IN
SELECT
COLUMN_WE_WANT
FROM
TABLE_B
WHERE
SOME_PERCENTAGE_RATE_COLUMN < 90
)There is a spatial forum. You will likely have a far better experience there.
-
Oracle Spatial Query from SQLPLUS deteriorates in performance
Hi,
I am running an sql-procedure, which does spatial queries. When I run this procedure (from SQLPLUS) for the first time it takes around 45 sec to complete, but when I run the same procedure for the second time it takes 65 seconds and the performance keeps on decreasing. But if I open a new SQLPLUS session it takes 45 secs and it also starts deteriorating subsequently. This strange behaviour is only for spatial queries, if I don't have spatial queries in my procedure then its fine.
I found this behaviour for any type of spatial query.
Is there something which can be done for improving it ? I found a lot of information on improving performance of a query, but my problem is to maintain the same performance of the same query in the same session.
Thanks,
RishiHi,
Could you provide version information?
Also, could you try monitoring memory as the procedure is running an see
if memory climbs more each time the procedure is run?
Thanks -
Spatial query on join view in another schema
Background:
I have a schema with tables containing four spatial columns per table. I have a second schema that contains join views of the tables in the first schema. The join view consists of the attributes and any other tabular data that is joined via foreign key to non-spatial tables and one of the four geometry columns. I have a third schema that owns no database objects of its own, but has select privilege to the second schema's view objects.
We are using Oracle 10.2.0.2 on a windows platform.
This worked in 9i, the problem only presented itself when we migrated to 10g.
Problem:
When I run a spatial query on a join view in the second schema(the schema that owns the object), it runs and returns results in about a second. When I run the exact same query in the third schema it routinely takes 20-30 seconds.
Here is the query that I run in both schemas with varying query times:
SELECT COUNT(*)
FROM two.join_view_a A
WHERE (MDSYS.SDO_FILTER(A.GEOM, SDO_GEOMETRY(2003,8307,NULL,SDO_ELEM_INFO_ARRAY(1,1003,3),SDO_ORDINATE_ARRAY(26.6396,26.5708,43.346,34.4083)),'QUERYTYPE=WINDOW')='TRUE');
Question:
Why does this happen?
What can I do to fix this problem?
Thanks,
JohnHave a read of this posting: Oracle spatial view runing slow as different user
There are details there about the 'optimizer_secure_view_merging' parameter and the effect it can have on cross schema views in 10.2 -
Spatial query in a distributed environment
We want to compare two spatial tables.
We have (at present identical) tables on two computers GEOSRV0 and NB25.
On both we have Oracle 10.1.0.2. GEOSRV0 has the WINDOWS 2000 operating system installed, on NB25 there is WINDOWS XP Professional.
The table is EZG:
GEOMETRY MDSYS.SDO_GEOMETRY
GEOMETRY_SK VARCHAR2(15)
EZG_NR NOT NULL NUMBER
FLAECHE FLOAT(126)
The primary key is EZG_NR.
There is a private database link for the user WLKREP on NB25 to user WLKREP on GEOSRV0.
The USER_SDO_GEOM_METADATA are the same on both machines (same SRID, same DIMINFO).
The following query â we check, if for each record on NB25 is a record on GEOSRV0 - works normally:
SQL> select ezg_nr from
2 ezg a where not exists
3 (select 1 from wlkrep.ezg@wlk where ezg_nr=a.ezg_nr);
no rows selected
This is correct.
We can read the geometry on the remote database:
SQL> select * from wlkrep.ezg@wlk where rownum=1;
GEOMETRY(SDO_GTYPE, SDO_SRID, SDO_POINT(X, Y, Z), SDO_ELEM_INFO, SDO_ORDINATES)
GEOMETRY_SK EZG_NR FLAECHE
SDO_GEOMETRY(3003, 1040001, NULL, SDO_ELEM_INFO_ARRAY(1, 1003, 1), SDO_ORDINATE_ARRAY(182285.489, 405672.889, 0, 182248.701, 405644.889, 0, 182241.984, 405639.771, 0, 182171.753, 405538.805, 0, 182130.953, 405498.016, 0, 181971.1, 405274.91
6, 0, 182223.796, 405176.609, 0, 182292.701, 405204.37, 0, 182439.028, 405451.66
, 0, 182573.506, 405628.153, 0, 182643.64, 405726.59, 0, 182718.945, 405764.608,
0, 182780.371, 405805.57, 0, 182862.191, 405843.931, 0, 182966.69, 405933.13, 0
, 183131.065, 406064.852, 0, 183205.098, 406160.637, 0, 183231.535, 406220.959,
0, 183266.473, 406265.415, 0, 183308.669, 406303.524, 0, 183295.305, 406364.696,
0, 183204.283, 406331.35, 0, 183161.835, 406299.691, 0, 183134.554, 406243.403,
GEOMETRY(SDO_GTYPE, SDO_SRID, SDO_POINT(X, Y, Z), SDO_ELEM_INFO, SDO_ORDINATES)
GEOMETRY_SK EZG_NR FLAECHE
0, 183080.454, 406171.548, 0, 183027.376, 406123.698, 0, 182914.558, 406039.291
, 0, 182770.335, 405973.138, 0, 182593.691, 405886.413, 0, 182443.506, 405795.85
2, 0, 182343.293, 405713.991, 0, 182285.489, 405672.889, 0))
13 304906.065
Now we want to compare the geometry and get
SQL> select a.ezg_nr from
2 wlkrep.ezg@wlk b, ezg a where a.ezg_nr=b.ezg_nr and
3 sdo_equal(a.geometry,b.geometry) != 'TRUE';
select a.ezg_nr from
ERROR at line 1:
ORA-03113: end-of-file on communication channel
Only comparing two elements gives the same result
SQL> select a.ezg_nr from
2 wlkrep.ezg@wlk b, ezg a where a.ezg_nr=640 and
3 b.ezg_nr = 640 and
4 sdo_equal(a.geometry,b.geometry) = 'TRUE';
select a.ezg_nr from
ERROR at line 1:
ORA-03113: end-of-file on communication channel
Even if we query a non existing element
select count(*) from ezg where ezg_nr=99999;
COUNT(*)
0
SQL> select a.ezg_nr from
2 wlkrep.ezg@wlk b, ezg a where a.ezg_nr=99999 and
3 b.ezg_nr = 99999 and
4 sdo_equal(a.geometry,b.geometry) = 'TRUE';
select a.ezg_nr from
ERROR at line 1:
ORA-03113: end-of-file on communication channel
Is this a bug or is a distributed spatial query not supported?
The following statement works:
SQL> select a.ezg_nr from
2 wlkrep.ezg@wlk a where not exists
3 (select 1 from ezg b where
4 b.ezg_nr=a.ezg_nr and
5 sdo_equal(b.geometry,a.geometry)='TRUE');
no rows selected
Regards
KarlActually, I have filed a TAR to Oracle on this. I also spoke to Steve Muench about it at OracleWorld, briefly. As of today, I'm trying to add to DataScroller.java to make it implement java.io.Serializable (with no luck...maybe someone can comment on what I'm doing wrong).
Here's what I have done so far.
Changed DataScroller.java (located in bc4jhtmlsrc.zip) to the following:
public class DataScroller implements java.io.Serializable
I added the following members per the docs on java.io.Serializable
private void writeObject(java.io.ObjectOutputStream out)
throws IOException
// really don't need to save state. Next call will recompute what's necessary
private void readObject(java.io.ObjectInputStream in)
throws IOException, ClassNotFoundException
// really don't need to load state. Next call will recompute what's necessary
I then used jar xvf to "unjar" datatags.jar, replaced DataScroller.class with my version and used
jar cv0Mf datatags.jar META-INF/* oracle/*
to recreate the datatags.jar. My file is not 100% the exact same as their file, but it looks to be right to me. I ended up doing it this way because the oracle\ord\* classes are not in the bc4jhtmlsrc.zip, only in the jar file and felt this was the "safest" way to include the correct contents of the file.
HOWEVER -- I'm sorry to say that I've then tried to run my app and get the same issue.
Thanks,
Jeff -
Spatial Query with multiple geometries from 2 tables
Hi,
I'm using Oracle 8.1.7. I am trying to do a spatial query on two tables with multiple geometries. There are two spatial tables. One made up of points and one made up of polygons. There are 829551 rows in the table of points and 1817795 rows in the table with polygons...
I want to find all polygons where one of this points is within it...
This query is pretty intensive querying two large spatial tables against each other and so I need to find the most efficient way of running this query. I've been running variations of my query for the last two week and every time it's bombed out with various errors after 12-24 hrs processing like out of memory, out of tablespace, out of processing, invalid index etc etc etc. I need to get this query run asap... Any tips would be gratefully appreciated..
For the session running the query I've allocated 16M sort area with
ALTER SESSION SET SORT_AREA_SIZE=16777216;
Below is the query I'm running... How can I improve this please? BTW PARCEL_OVERLAPS contains the points and TP$_PARCEL_AREAS the polygons.
SELECT lu.LNU_PARCEL_ID
FROM
seventy.PARCEL_OVERLAPS po,
imap_topol.TP$_PARCEL_AREAS pa,
TP$_PARCEL_CENTROID_AREA pca,
TDCR_LAND_UNIT lu
WHERE
SDO_FILTER(po.geometry, pa.area,
'querytype=WINDOW') = TRUE and
sdo_within_distance(po.geometry,pa.area,'distance=0')='TRUE' and
pa.delete_date is null and
Lu.LNU_LAND_UNIT_UNIQUE_NUM = pca.CENTROID_ID and
pa.AREA_ID = pca.AREA_ID and
pca.DELETE_DATE IS NULL and
pa.DELETE_DATE IS NULL;Albert,
Thanks for your reply and the tips you've given. The tp$_parcel_areas table will always be bigger so I've changed the order to sdo_within_distance(pa.area,po.geometry,'distance=0')='TRUE'. The requested counts for those queries are
12:26:29 [email protected]:SQL> select count(*)
13:46:22 2 from seventy.PARCEL_OVERLAPS;
COUNT(*)
612
13:48:12 [email protected]:SQL> select count(*)
13:48:17 2 from imap_topol.TP$_PARCEL_AREAS pa,
13:48:21 3 TP$_PARCEL_CENTROID_AREA pca
13:48:21 4 where pca.DELETE_DATE IS NULL
13:48:21 5 and pa.DELETE_DATE IS NULL
13:48:21 6 and pa.AREA_ID = pca.AREA_ID;
COUNT(*)
1310665
There was no reason for both filter and within_distance. I did try use the anyinteract but for some reason that does not return the desired results(I've added one id row as a test to make sure it returns desired results). Plus Oracle have recomended using the within distance for better performance..
so the explan plan for
14:38:37 [email protected]:SQL> EXPLAIN PLAN FOR
14:38:50 2 SELECT lu.LNU_PARCEL_ID
14:38:50 3 FROM
14:38:50 4 seventy.PARCEL_OVERLAPS po,
14:38:50 5 imap_topol.TP$_PARCEL_AREAS pa,
14:38:50 6 TP$_PARCEL_CENTROID_AREA pca,
14:38:50 7 TDCR_LAND_UNIT lu
14:38:50 8 WHERE
14:38:50 9 sdo_within_distance(pa.area,po.geometry,'distance=0')='TRUE' and
14:38:50 10 pa.delete_date is null and
14:38:50 11 Lu.LNU_LAND_UNIT_UNIQUE_NUM = pca.CENTROID_ID and
14:38:50 12 pa.AREA_ID = pca.AREA_ID and
14:38:50 13 pca.DELETE_DATE IS NULL and
14:38:50 14 pa.DELETE_DATE IS NULL;
is
Plan Table
| Operation | Name | Rows | Bytes| Cost | Pstart| Pstop |
| SELECT STATEMENT | | 4G|32920G| 547M| | |
| NESTED LOOPS | | 4G|32920G| 547M| | |
| MERGE JOIN | | 547M| 2029G| 230124 | | |
| SORT JOIN | | 1M| 36M| 85014 | | |
| MERGE JOIN | | 1M| 36M| 50019 | | |
| SORT JOIN | | 1M| 17M| 21650 | | |
| TABLE ACCESS FULL |TP$_PARCE | 1M| 17M| 485 | | |
| SORT JOIN | | 1M| 22M| 28369 | | |
| TABLE ACCESS FULL |TDCR_LAND | 1M| 22M| 2127 | | |
| SORT JOIN | | 42K| 160M| 145111 | | |
| TABLE ACCESS FULL |TP$_PARCE | 42K| 160M| 12697 | | |
| TABLE ACCESS FULL |PARCEL_OV | 817 | 3M| 1 | | |
14:43:14 [email protected]:SQL> explain plan for
14:43:23 2 SELECT pa.AREA_ID
14:43:23 3 FROM seventy.PARCEL_OVERLAPS po,
14:43:23 4 imap_topol.TP$_PARCEL_AREAS pa
14:43:23 5 WHERE SDO_RELATE(po.geometry, pa.area,'mask=ANTINTERACT querytype=WINDOW') = 'TRUE'
14:43:23 6 and pa.DELETE_DATE IS NULL;
Plan Table
| Operation | Name | Rows | Bytes| Cost | Pstart| Pstop |
| SELECT STATEMENT | | 6M| 50G| 10M| | |
| NESTED LOOPS | | 6M| 50G| 10M| | |
| TABLE ACCESS FULL |PARCEL_OV | 817 | 3M| 1 | | |
| TABLE ACCESS FULL |TP$_PARCE | 850K| 3G| 12697 | | |
14:45:03 [email protected]:SQL> explain plan for
14:45:04 2 SELECT pa.AREA_ID
14:45:05 3 FROM seventy.PARCEL_OVERLAPS po,
14:45:05 4 imap_topol.TP$_PARCEL_AREAS pa
14:45:05 5 WHERE SDO_RELATE(pa.area, po.geometry,'mask=ANTINTERACT querytype=WINDOW') = 'TRUE'
14:45:05 6 and pa.DELETE_DATE IS NULL;
Plan Table
| Operation | Name | Rows | Bytes| Cost | Pstart| Pstop |
| SELECT STATEMENT | | 6M| 50G| 863554 | | |
| NESTED LOOPS | | 6M| 50G| 863554 | | |
| TABLE ACCESS FULL |TP$_PARCE | 850K| 3G| 12697 | | |
| TABLE ACCESS FULL |PARCEL_OV | 817 | 3M| 1 | | |
-------------------------------------------------------------------------------- -
Export with embedded spatial query
I am trying to use the exp utility to export spatial data. I am using the "query" parameter to specify a spatial query based on another spatial layer.
I have two layers. Admin1 has lots of administrative boundaries. PE has three polygons in it over different sections of my data set. I want to get the data extracted from admin1 that exists in these areas to an export file(along with all the triggers and other related database objects). I was hoping this approach would work
I am getting errors saying that the identifier is too long.
Here is the par file:
file=1.exp
log=1.log
tables=(admin1, pe)
query="""WHERE mdsys.SDO_RELATE(GEOM1,pe.GEOM, 'MASK=INSIDE QUERYTYPE=WINDOW')='TRUE' OR SDO_RELATE(GEOM2,pe.GEOM, 'MASK=INSIDE QUERYTYPE=WINDOW')='TRUE'"""
I get:
Export: Release 10.2.0.1.0 - Production on Wed May 10 11:56:14 2006
Copyright (c) 1982, 2005, Oracle. All rights reserved.
Connected to: Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - Produc
tion
With the Partitioning, OLAP and Data Mining options
Export done in WE8MSWIN1252 character set and AL16UTF16 NCHAR character set
About to export specified tables via Conventional Path ...
. . exporting table ADMIN1
EXP-00056: ORACLE error 972 encountered
ORA-00972: identifier is too long
. . exporting table PE
EXP-00056: ORACLE error 972 encountered
ORA-00972: identifier is too long
Export terminated successfully with warnings.
Is this even possible? Can you use the exp query parameter to perform a spatial operation? Did this work? I am accustom to seeing record counts next to the exported tables, and they are not present in the output from exp. The export file is not empty. It exported all the features, not the subset I had requested.
How can I structure the query differently so that I get the export file with just the lines that fall within the polygons?
Thanks,
JohnI am trying to use the exp utility to export spatial data. I am using the "query" parameter to specify a spatial query based on another spatial layer.
I have two layers. Admin1 has lots of administrative boundaries. PE has three polygons in it over different sections of my data set. I want to get the data extracted from admin1 that exists in these areas to an export file(along with all the triggers and other related database objects). I was hoping this approach would work
I am getting errors saying that the identifier is too long.
Here is the par file:
file=1.exp
log=1.log
tables=(admin1, pe)
query="""WHERE mdsys.SDO_RELATE(GEOM1,pe.GEOM, 'MASK=INSIDE QUERYTYPE=WINDOW')='TRUE' OR SDO_RELATE(GEOM2,pe.GEOM, 'MASK=INSIDE QUERYTYPE=WINDOW')='TRUE'"""
I get:
Export: Release 10.2.0.1.0 - Production on Wed May 10 11:56:14 2006
Copyright (c) 1982, 2005, Oracle. All rights reserved.
Connected to: Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - Produc
tion
With the Partitioning, OLAP and Data Mining options
Export done in WE8MSWIN1252 character set and AL16UTF16 NCHAR character set
About to export specified tables via Conventional Path ...
. . exporting table ADMIN1
EXP-00056: ORACLE error 972 encountered
ORA-00972: identifier is too long
. . exporting table PE
EXP-00056: ORACLE error 972 encountered
ORA-00972: identifier is too long
Export terminated successfully with warnings.
Is this even possible? Can you use the exp query parameter to perform a spatial operation? Did this work? I am accustom to seeing record counts next to the exported tables, and they are not present in the output from exp. The export file is not empty. It exported all the features, not the subset I had requested.
How can I structure the query differently so that I get the export file with just the lines that fall within the polygons?
Thanks,
John -
hi, there.
I am working Oracle Spatial 9.0.1 and I have a problem about the
performance of spaital query, can someone help me resolve it?
/*query 1*/
select t1.line_seg
from spatial_list_Traj
where sdo_within_distance(convert_trajseg_to_lineseg
(t1.line_seg), MDSYS.SDO_GEOMETRY(2003, NULL, NULL,
MDSYS.SDO_ELEM_INFO_ARRAY(1, 1003, 3), MDSYS.SDO_ORDINATE_ARRAY
(212.117, 352.613, 221.617, 362.708)), 'querytype=filter
distance=0.5') = 'TRUE';
/* query 2*/
select t1.line_seg
from spatial_list_traj
where object_id = 1 and sdo_within_distance
(convert_trajseg_to_lineseg(t1.line_seg), MDSYS.SDO_GEOMETRY
(2003, NULL, NULL, MDSYS.SDO_ELEM_INFO_ARRAY(1, 1003, 3),
MDSYS.SDO_ORDINATE_ARRAY(212.117, 352.613, 221.617,
362.708)), 'querytype=filter distance=0.5') = 'TRUE';
there are about 120,000 rows in the spatial_list_Traj table and
query 1 takes only 1 second to finish, which is a good result
but if I add another condition, like query 2, the time cost is
still 1 second. I thought it can be much faster since there is
only 48 rows whose object_id is 1. what I expected it to do was
for the optimizer to filter out the rows that match
the 'object_id = 1' criteria and then do a spatial query on the
filtered data.
PS: there are two indices--a b-tree index on column 'object_id'
and of course, a functional spatial index on column 'line_seg',
which is a R-tree.
thanks in advance
Hu CaoHi Hu,
The second query is probably not using the spatial index. The
spatial index indexes the whole table based on spatial
relationships, so even though you may have filtered out all but
48 rows, the spatial index has to be applied to all of the rows.
In this case, it is probable that all of the rows returned with
object_id=1 are being geometrically compared with your query
window, which may account for the extra time.
You may want to try a hint to tell the optimizer not to use the
index on object_id, i.e.
SELECT /*+NO_INDEX(spatial_list_traj object_id_index_name)*/
Hope this helps,
dan -
Simple Spatial Query caused ORA-03113: end-of-file on communication channel
I try to run a simple spatial query to find the ID's which inside the specified circle
select n.ID,n.FEATURE_ID
from NC_MANGROVE n
where sdo_relate(n.shape,
mdsys.sdo_geometry(2003,null,null,
mdsys.sdo_elem_info_array(1,1003,4),
mdsys.sdo_ordinate_array(848948,834940,848908,834980,848908,834900)),
'mask=ANYINTERACT querytype=WINDOW')='TRUE';
but it end up of ORA-03113: end-of-file on communication channel
Then I try another query, which I think do the similar thing
select n.ID,n.FEATURE_ID
from NC_MANGROVE n
where sdo_within_distance(n.shape,
mdsys.sdo_geometry(2001,null,mdsys.sdo_point_type(848908,834900,null),null,null),
'distance=40')='TRUE';
but same error occurs.
Is it the SQL*net configure problem. I am using Oracle 8.1.7 for both client and server.
Please help
TommyDoes SDO_FILETER Queries work fine? If so there is already reported bug and patch for that on Metalink site. Get all the patches for 8.17 spatial and install those.
Maybe you are looking for
-
MacBook Pro (Early 2011) behaves strange after GPU replacement
I have an early MBP 2011 and had the GPU issue. After Apple finally announced the replace program I immediatelly went to a a stor and the fixed it. But now my System acts very strangely ever since. Here a list of all the strange things: - The overall
-
I want to put a 3rd party SSD in my mac mini and I no how to put in inside the mac mini but, how do I get osx on it preferably Yosemite I do not have a external hard drive just the one that comes with the mac mini and my 3rd party internal SSD please
-
ipad 2 screen is black . when i press the power button it takes a screen shot but i have no functionality . tried holding the home and power button together but the screen just flicker and stay black . no apple icon shows up. help please
-
Sap BW Statistic Report - 0TCTUTIME key figure
Hi Im trying to have a report avaiable to display the # of hits per user for a given BW Bex report. The proposed solution is to display the # of hits by using characteristic 0TCTUTIME. Is this the right characteristic to use to the diplay the # of hi
-
Change of Depreciation key in SAP as per NEW Co Act 2013
Dear Expert, As per New Companies Act 2013, we need to change the depreciation key in the SAP FI Module, kindly suggest the process how it need to do. Thanks in Advance..!!