Hello, I noticed when adding a column of type 'POINTZ' with SRID in

<a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/us

Conversion from EWKB does not support 3d about geomet HOT 1 OPEN

geomet commented on June 8, 2024

Conversion from EWKB does not support 3d

from geomet.

Comments (1)

larsbutler commented on June 8, 2024 1

@frisia-mtz Thanks for the bug report. I appreciate you taking the time to point this out and provide an example.

I think this is more of a general problem of the way the geometry type (bytes 1-5: 010000A0) is represented in EWKB context. In EWKB, the 01 indicates a Point-type, and the A0 is the bitwise OR [1] of an SRID flag (0x20) and a WKB Z offset (0x80) [2] --> 0x20 | 0x80 = 0xa. geomet is checking for the presence of the SRID flag [3] but it's not expecting the dimensional offset.

It looks the reverse of the conversion (dumps/to WKB) is also not correctly setting the dimensionality flag [4].

It looks like the correct fix would be to:

look for the dimensionality flag when parsing EWKB
inject the dimensionality flag when generating EWKB

@tomplex @achapkowski Would you both agree with this assessment? Is there anything I'm missing, perhaps an alternate EWKB spec independent of PostGIS that would contradict this and cause additional breakage in other use cases, if we were to fix this per the PostGIS spec?

References:
[1] - https://github.com/postgis/postgis/blob/b1b8c11bfe54d7756ed53719e74783dd2733d3df/doc/ZMSgeoms.txt#L56-L66
[2] - https://github.com/postgis/postgis/blob/b1b8c11bfe54d7756ed53719e74783dd2733d3df/liblwgeom/liblwgeom.h.in#L124-L127
[2] -

geomet/geomet/wkb.py

Line 135 in 8d2b8f2

has_srid = high_byte == b'\x20'

[3] -

geomet/geomet/wkb.py

Lines 336 to 373 in 8d2b8f2

 def _header_bytefmt_byteorder(geom_type, num_dims, big_endian, meta=None): 

 """ 

  Utility function to get the WKB header (endian byte + type header), byte 

  format string, and byte order string. 

  """ 

 dim = _INT_TO_DIM_LABEL.get(num_dims) 

 if dim is None: 

 pass # TODO: raise 

 type_byte_str = _WKB[dim][geom_type] 

 srid = meta.get('srid') 

 if srid is not None: 

 # Add the srid flag 

 type_byte_str = SRID_FLAG + type_byte_str[1:] 

 if big_endian: 

 header = BIG_ENDIAN 

 byte_fmt = b'>' 

 byte_order = '>' 

 else: 

 header = LITTLE_ENDIAN 

 byte_fmt = b'<' 

 byte_order = '<' 

 # reverse the byte ordering for little endian 

 type_byte_str = type_byte_str[::-1] 

 header += type_byte_str 

 if srid is not None: 

 srid = int(srid) 

 if big_endian: 

 srid_header = struct.pack('>i', srid) 

 else: 

 srid_header = struct.pack('<i', srid) 

 header += srid_header 

 byte_fmt += b'd' * num_dims 

 return header, byte_fmt, byte_order

from geomet.

Conversion from EWKB does not support 3d about geomet HOT 1 OPEN

Comments (1)

Related Issues (20)

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent

	def _header_bytefmt_byteorder(geom_type, num_dims, big_endian, meta=None):
	"""
	Utility function to get the WKB header (endian byte + type header), byte
	format string, and byte order string.
	"""
	dim = _INT_TO_DIM_LABEL.get(num_dims)
	if dim is None:
	pass # TODO: raise

	type_byte_str = _WKB[dim][geom_type]
	srid = meta.get('srid')
	if srid is not None:
	# Add the srid flag
	type_byte_str = SRID_FLAG + type_byte_str[1:]

	if big_endian:
	header = BIG_ENDIAN
	byte_fmt = b'>'
	byte_order = '>'
	else:
	header = LITTLE_ENDIAN
	byte_fmt = b'<'
	byte_order = '<'
	# reverse the byte ordering for little endian
	type_byte_str = type_byte_str[::-1]

	header += type_byte_str
	if srid is not None:
	srid = int(srid)

	if big_endian:
	srid_header = struct.pack('>i', srid)
	else:
	srid_header = struct.pack('<i', srid)
	header += srid_header
	byte_fmt += b'd' * num_dims

	return header, byte_fmt, byte_order