CODEWITHSUNDEEP
pythonoracle-databaseoracle-spatialpython-oracledb
zedaardv
zedaardv

Reputation: 1

Python dynamic processing of Oracle SDO_GEOMETRY objects

There are some situations where the WKT format does not work (3D,4D geometries, arc in a geodetic). These are handled in Postgis EWKT format.

I have an out converter

def SDOOutConverter(DBobj):
    
  if DBobj.SDO_ELEM_INFO:
    sdo_elem = DBobj.SDO_ELEM_INFO.aslist()
  else:
    sdo_elem = None
      
  if DBobj.SDO_ORDINATES:
    sdo_ord = DBobj.SDO_ORDINATES.aslist()
  else:
    sdo_ord = None 
  return lSDO( 
    int(DBobj.SDO_GTYPE)
    ,DBobj.SDO_SRID
    ,sdo_elem
    ,sdo_ord
  )

and the object definition.

class lSDO(object):
  def __init__(self, gtype, srid, elemInfo, ordinates):
    self.gtype = gtype
    self.srid = int(srid or 0)
    self.elemInfo = elemInfo
    self.ordinates = ordinates

I have an example of returning the object information from a fixed query

cur.execute('select  mkt_id, name, shape from hamster.cola_markets t')
for time_id, name, obj in cur:
      sdo_type = obj.gtype
      sdo_srid = obj.srid
      sdo_elem = obj.elemInfo
      sdo_ord = obj.ordinates

This takes the geometry column (shape) then in the for loop access the object.

Question: How would i do this dynamically. There are several tables with differing structures.

how would I write a dynamic for time_id, name, obj in cur:?

bonus question, How would one handle a table with multiple GEOMETRY columns?

I can crate a column list and tried using the string in the for loop

for '{}'.format(objcol) in cur:

results in:

   for '{}'.format(objcol) in cur:
        ^^^^^^^^^^^^^^^^^^^
SyntaxError: cannot assign to function call

The problem I'm trying to solve is: replacing

cur.execute('select  mkt_id, name, shape from hamster.cola_markets t')
for time_id, name, obj in cur:

with:

cur.execute('select * from {}.{}'.format(schema,tablename))
for [dynamic column list with obj reference] in cur:

In other words handle cursor procession dynamically.

Upvotes: 0

Views: 215

Answers (1)

zedaardv
zedaardv

Reputation: 1

I'm going to partially answer my own question.

Part of my confusion is from the examples for sdo_geometries, without clear explanations.

With further experimenting i discovered that

cur.execute('select  mkt_id, name, shape from hamster.cola_markets t')
for time_id, name, obj in cur:

is misleading. obj is just an alias for shape, so

sdo_type = obj.gtype

would work as well as

sdo_type = shape.gtype

So the problem for me was attributing some special meaning to obj, when what is does, at least for me, is obfuscates what is actually going on.

Documenting what is going on in code is helpful.

Upvotes: 0

Related Questions