"Supplied map is missing required keys" when using enlarge

Community
1

I am trying to use enlarge to buffer a dataset of points but it is failing with this error

riskscape pipeline eval pipeline_buffer_points.txt
[WARNING] The 'beta' plugin is enabled. This contains experimental features that may significantly change or be deprecated in future releases.
Problems found with 'pipeline-eval' pipeline
  - Execution of your data processing pipeline failed. The reasons for this follow:
    - Supplied map '{...}' is missing required keys - [INSP_DATE, FLOOD_FOUNDATION_TYPE, FLOOD_PERIMETER, FLOOD_HEIGHT_CONF, FLOOD_NOT_MEAS_RSN, FLOOD_STOREYS, FLOOD_UNIT_LEVEL, FLOOD_CLADDING_TYPE, FLOOD_COMMENT, FLOOD_FRONT_SILL_HEIGHT, FLOOD_OTHER_SILL_HEIGHT]

my pipeline is

input(relation: 'points', name: 'exposure') as exposures_input
 -> enlarge(distance: 10, mode: 'ROUND', remove-overlaps: false)
 -> save(name:'buffered')

If I comment out the enlarge step then it runs fine. It also runs fine if I use a trivial dataset of points with no attributes

2

Hi John,

That error message isn’t particularly helpful, is it? Could you send through the output from the following command?

riskscape pipeline eval pipeline_buffer_points.txt --print

That might give a few more clues as to what’s happening here.

In the meantime, you could also try doing the buffering long-hand, using something like this:

input(relation: 'points', name: 'exposure') as exposures_input
 -> select({ merge(exposure, { buffer(exposure.the_geom, 10) as the_geom }) as exposure })
 -> save(name:'buffered')

You’d need to replace the_geom with the actual name of the geometry attribute in your case.

Cheers,
Tim

3 
$ riskscape pipeline eval pipeline_buffer_points.txt --print
[WARNING] The 'beta' plugin is enabled. This contains experimental features that may significantly change or be deprecated in future releases.
Step: exposures_input:[input]
  Parameters:
    limit : <none-given>
    name : exposure
    offset : <none-given>
    relation : FeatureSourceRelation
    value : <none-given>
  Produces:
      exposure =>
        geom => Point[crs=EPSG:2193]
        OBJECTID => Integer
        EXTRACT_DATE => Date
        LETTABLE_UNIT_CODE => Text
        FORMATTED_ADDRESS => Text
        PROPERTY_TYPE_CODE => Text
        PROPERTY_TYPE_DESC => Text
        PROPERTY_OWNER_CODE => Text
        ASSET_TYPE_DESC => Text
        PROPERTY_DESCRIPTION => Text
        CURR_TLA => Text
        CITY => Text
        PROPERTY_LAND_CODE => Text
        LAND_STATUS => Text
        NUMBER_OF_BEDROOMS_CNT => Integer
        BUILD_YEAR => Integer
        BLDG_NUMBER_OF_STOREYS => Integer
        CLADDING_TYPE => Text
        LAND_AREA_SQM => Integer
        PROPERTY_FLOOR_AREA_SQM => Floating
        Operating_Region => Text
        Parcel_ID => Integer
        INSP_DATE => Date
        FLOOD_FOUNDATION_TYPE => Text
        FLOOD_PERIMETER => Text
        FLOOD_HEIGHT_CONF => Text
        FLOOD_NOT_MEAS_RSN => Text
        FLOOD_STOREYS => Integer
        FLOOD_UNIT_LEVEL => Text
        FLOOD_CLADDING_TYPE => Text
        FLOOD_COMMENT => Text
        FLOOD_FRONT_SILL_HEIGHT => Floating
        FLOOD_OTHER_SILL_HEIGHT => Floating
        Final_valuation => Floating
  Result: FeatureSourceRelation
Step: save:[save]
  Inputs: [exposures_input:[input]]
  Parameters:
    format : <none-given>
    name : buffered
  Produces:
  Result: SaveSinkConstructor
4

Okay sweet the long hand way is running now. Thanks

5

That’s good. The problem may have been that some of those attributes are null in the input data. You could try running something like this:

riskscape pipeline eval "input('points') -> filter(is_null(INSP_DATE))"

Then check the output - if there are features in it, then that means some of the attributes in the input data are null.

If that’s the case, then you can probably fix the enlarge() problem by using a bookmark type - either by marking the attributes as being ‘nullable’ type, or just ignoring the attributes if you don’t need them in the output.