| Ticket Hash: | 8c09a2fd879dc4e7d5f0679f7a62f2dddce7d4bf | |||
| Title: | ST_Transform should ignore Z value as default, adding bAttempt3D Parameter | |||
| Status: | Tested | Type: | Code_Defect | |
| Severity: | Critical | Priority: | Immediate | |
| Subsystem: | Resolution: | Fixed | ||
| Last Modified: | 2017-09-25 13:32:58 | |||
| Version Found In: | dev | |||
| User Comments: | ||||
|
mj10777 added on 2017-09-10 05:19:46:
While importing the 1.116 billion 1 meter based height data for Berlin,
Sample: echo "391475.000 5819347.000 30.11" | cs2cs +init=epsg:25833 +to +init=epsg:3068 24747.12 20725.15 -10.81A difference of 40.92 meters. In Berlin the range is from 24.0 to 120.0 meters. Making the results completely unusable. For this reason I would suggest changing the default behavior of ST_Transform (gaiaTransformCommon), so that the z values are not given to pj_transform
/* applying reprojection */
if (is2d)
rc=pj_transform (from_cs, to_cs, cnt, 0, xx, yy, NULL);
else
rc=pj_transform (from_cs, to_cs, cnt, 0, xx, yy, zz);
if (rc == 0)
{
// needed 4 times inside gaiaTransformCommon
Since this is a destructive function, attempting to utilize a 3D transformation should be made optional. A clear warning in the present documentation and for a new optional parameter should be clearly stated. I have no idea what proj4 does here, but if they don't take into consideration that each system could have a different '0' point as it's base (which in time could change), a falsification is very likely. I hard-coded a 'int is2d=1' in gaiaTransformCommon, after which the above result came out correctly. gaiaTransformCommon is called from:
gaiaTransform_r
fnct_AsKml1 fnct_AsKml3 mj10777 added on 2017-09-25 13:32:58: The implemented new Function 'ST_TransformXY()' works as expected. | ||||