Make __id__ column optional (it's required now)

[pnorman]

When I try to serve data without an __id__ column I get the error

192.168.1.2 - - [09/Jun/2017 21:10:14] "GET /all/0/0/0.json HTTP/1.1" 500 -
Traceback (most recent call last):
  File "/home/pnorman/osm/tilezen/tileserver/tileserver/__init__.py", line 148, in __call__
    response = self.handle_request(request)
  File "/home/pnorman/osm/tilezen/tileserver/tileserver/__init__.py", line 232, in handle_request
    self.post_process_data,
  File "/home/pnorman/osm/tilezen/tileserver/env/lib/python2.7/site-packages/tilequeue/process.py", line 243, in process_coord_no_format
    feature_id = row.pop('__id__')
KeyError: '__id__'

To get around this I added the column 0::bigint AS __id__, but I shouldn't need to do this. Not all data sources have meaningful IDs, or you might not want to include them in the output.

My queries file is

all:
  - water
layers:
  water:
    template: water.jinja2
    start_zoom: 0
    geometry_types: [Polygon]

and water.jinja2 is

SELECT
    {% filter geometry %}{{ bounds['polygon']|bbox_padded_intersection('way') }}{% endfilter %} AS __geometry__,
    0::bigint AS __id__
  FROM ne_ocean
  WHERE {{ bounds['polygon']|bbox_filter('way',3857) }}

[nvkelso]

We drop some ids in post-processing, so this seems reasonable.

…

On Jun 9, 2017, at 21:18, Paul Norman ***@***.***> wrote: When I try to serve data without an __id__ column I get the error 192.168.1.2 - - [09/Jun/2017 21:10:14] "GET /all/0/0/0.json HTTP/1.1" 500 - Traceback (most recent call last): File "/home/pnorman/osm/tilezen/tileserver/tileserver/__init__.py", line 148, in __call__ response = self.handle_request(request) File "/home/pnorman/osm/tilezen/tileserver/tileserver/__init__.py", line 232, in handle_request self.post_process_data, File "/home/pnorman/osm/tilezen/tileserver/env/lib/python2.7/site-packages/tilequeue/process.py", line 243, in process_coord_no_format feature_id = row.pop('__id__') KeyError: '__id__' To get around this I added the column 0::bigint AS __id__, but I shouldn't need to do this. Not all data sources have meaningful IDs, or you might not want to include them in the output. My queries file is all: - water layers: water: template: water.jinja2 start_zoom: 0 geometry_types: [Polygon] and water.jinja2 is SELECT {% filter geometry %}{{ bounds['polygon']|bbox_padded_intersection('way') }}{% endfilter %} AS __geometry__, 0::bigint AS __id__ FROM ne_ocean WHERE {{ bounds['polygon']|bbox_filter('way',3857) }} ― You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.