search

GeoLatte / geolatte-geom

A geometry model that conforms to the OGC Simple Features for SQL specification.
Other
132 stars 63 forks source link

GEOMETRYCOLLECTION with MULTIPOINT from Oracle -> can't parse. #140

Closed Kermit closed 2 years ago

Kermit commented 2 years ago

Hi.

I have a problem with newest Geolatte used with Oracle 11g/12c. Geolatte is not able to properly parse GEOMETRYCOLLECTION with MULTIPOINT.

This is WKT of problematic geometry (SRID-2180): GEOMETRYCOLLECTION (LINESTRING (210949.01 638433.59, 210957.0 638454.21), LINESTRING (210945.64 638407.43, 210945.64 638407.44, 210947.6 638421.03, 210947.6 638421.04), LINESTRING (210932.36 638722.29, 210932.37 638722.29), LINESTRING (210921.9 638188.45, 210924.12 638190.58), MULTIPOINT ((210921.03 638080.97), (210919.92 638274.15), (210924.77 638056.98), (210929.6 638283.58), (210926.94 638271.23), (210930.11 637940.3), (210963.71 638468.55), (210955.13 637928.58), (210983.74 638504.63)), POLYGON ((211015.1 638577.87, 211013.45 638567.87, 211015.11 638577.88, 211015.1 638577.87)), POLYGON ((210996.51 638532.76, 211002.12 638533.0, 211002.123782194 638533.009723407, 210996.51 638532.76)), POLYGON ((210944.92 638429.26, 210949.0 638433.57, 210949.01 638433.59, 210944.92 638429.26)), POLYGON ((210942.64 637949.64, 210942.628783383 637949.63, 210942.64 637949.63, 210946.21 637952.83, 210946.209233244 637952.84137355, 210942.64 637949.64)), POLYGON ((210943.86 638387.06, 210942.541867683 638381.725000005, 210943.87 638387.05, 210945.64 638407.43, 210943.86 638387.06)), POLYGON ((210930.16 638304.53, 210929.877514327 638293.775, 210930.17 638304.52, 210935.07 638349.32, 210937.348586786 638354.020706598, 210935.06 638349.33, 210930.16 638304.53)), POLYGON ((210928.789310344 638001.199999998, 210929.13 638002.34, 210930.06 638003.99, 210929.13 638002.37, 210928.789310344 638001.199999998)), POLYGON ((210919.09 638012.53, 210921.82 638013.6, 210924.17 638017.48, 210928.53 638019.09, 210928.53 638019.11, 210924.18 638017.51, 210921.83 638013.64, 210919.09 638012.55, 210919.09 638012.53)), POLYGON ((210919.655908553 638009.779882999, 210919.09 638012.53, 210919.65 638009.78, 210919.655908553 638009.779882999)), POLYGON ((210920.659442823 638090.657599327, 210917.02 638089.83, 210917.018993718 638089.824307319, 210920.66 638090.65, 210920.659442823 638090.657599327)), POLYGON ((210914.7 638133.41, 210914.700075884 638133.404536378, 210916.76 638133.87, 210916.77 638133.88, 210914.7 638133.41)), POLYGON ((210874.328514051 637438.627868085, 210880.58 637436.66, 210888.36 637438.49, 210890.58 637449.85, 210906.61 637448.88, 211087.4 637405.78, 211158.93 637362.37, 211328.59 637334.07, 211301.49 637759.84, 211095.17 637803.82, 211093.0 637810.37, 211086.36 637810.48, 211078.75 637809.42, 211072.74 637815.45, 211072.81 637819.26, 211066.67 637821.9, 211055.85 637828.65, 211050.89 637833.94, 211052.18 637839.01, 211049.41 637840.99, 211045.72 637841.41, 211042.35 637841.98, 211042.33 637847.52, 211033.2 637858.53, 211037.64 637870.03, 211050.7 637879.49, 211055.32 637888.27, 211056.15 637892.62, 211058.39 637896.18, 211061.69 637897.71, 211063.75 637898.86, 211041.24 637905.58, 211027.25 637902.11, 211022.73 637901.74, 211018.734940562 637902.50234226, 211020.55 637901.81, 211028.66 637898.73, 211023.86 637887.92, 210874.328514051 637438.627868085)))

Shortened stacktrace: Can't determine ElementType from etype:1 and interp.:9 at org.geolatte.geom.codec.db.oracle.ElementType.parseType(ElementType.java:124) ~[geolatte-geom-1.9.0-SNAPSHOT.jar:na]

I found that, according to https://docs.oracle.com/database/121/SPATL/sdo_geometry-object-type.htm#GUID-270AE39D-7B83-46D0-9DD6-E5D99C045021__BGHDGCCE, SDO_INTERPRETATION with MULTIPOINT (cluster) informs us how many points geometry has. I'm assuming that this means its compound. In the ElementType.java file POINT_CLUSTER is described as not compound.

I tried to fix it by my own and I've created PR (#141) which I hope will be helpful. It works in my case.

Let me know here or in the PR (#141) if I have to provide some additional information.

Kermit commented 2 years ago

@maesenka what do you think about this?

maesenka commented 2 years ago

Hi @Kermit, first of all thank you for logging this issue and proposing a solution.

The Oracle SDO code is quite old so very possibly that some later addition (after Oracle 9c) haven't yet made it into the code. I will need some time to look into it, and properly assess your PR. But I will try to take this up as soon as possible.

Kermit commented 2 years ago

@maesenka thank you very much. I'll test it with my code and let you know if everything is ok!