Adding new attribute field to processing output algorithm in PyQGIS?

Question

I have a processing script that will carry out a selection by expression of the input layer, then do a join attributes by location with a join layer.

The next stage will be to add a new field and calculate the field, I got a python script that runs this but I am struggling to adapt it into the processing tool.

The code is below:

def initAlgorithm(self, config=None):
    """
    Here we define the inputs and output of the algorithm, along
    with some other properties.
    """

    # We add the input vector features source. It can have any kind of
    # geometry.
    self.addParameter(
        QgsProcessingParameterFeatureSource(
            self.INPUT,
            self.tr('NRD Layer'),
            [QgsProcessing.TypeVectorPoint]
        )
    )
    self.addParameter(
        QgsProcessingParameterFeatureSource(
            self.JOIN,
            self.tr('Parliamentary Constituency Layer'),
            [QgsProcessing.TypeVectorPolygon]
        )
    )
    self.addParameter(
        QgsProcessingParameterField(
            self.JOIN_FIELDS,
            self.tr('Choose Join Field'),
            parentLayerParameterName = self.JOIN,
            allowMultiple=False,
            optional=False
        )
    )
    self.addParameter(
        QgsProcessingParameterVectorDestination(
            self.OUTPUT,
            self.tr('Output layer')
        )
    )
def processAlgorithm(self, parameters, context, feedback):
    selectData = processing.run(
        'qgis:selectbyexpression',
        {
            "INPUT":parameters['INPUT'],
            "EXPRESSION":'\"mcmcode\" !=\'1\' AND \"mcmcode\" !=\'9\' AND \"mcmcode\" !=\'960\' AND \"mcmcode\" !=\'999\'',
            "METHOD":0,
            'OUTPUT':parameters['OUTPUT']
        },
        context=context,
        feedback=feedback 
    )
    joinoutput = processing.run(
    'qgis:joinattributesbylocation', 
    {
         'INPUT':QgsProcessingFeatureSourceDefinition(selectData['OUTPUT'], True),
         'JOIN':parameters['JOIN'] ,
         'PREDICATE':[0],
         'JOIN_FIELDS':parameters['join_field'],
         'METHOD':1,
         'DISCARD_NONMATCHING':True,
         'PREFIX':'',
         'OUTPUT':parameters['OUTPUT']
    }, 
    context = context, 
    feedback = feedback
    )
    NRDLayer = joinoutput['OUTPUT']
    F_Onset = QgsField('F_Onset',QVariant.Double)
    NRDLayer.dataProvider().addAttributes([F_Onset])
    NRDLayer.updateFields()

    return {'OUTPUT':joinoutput['OUTPUT']}

The error suggests you can't use dataprovider on joinoutput['OUTPUT'].

I thought I might need to use ParameterasSink to get my output in order to be able to use things like .featureCount(), .dataProvider() with it etc.

However not understanding this very well I would rather avoid that. The tool works up to that point.

Any ideas on what to do in QGIS 3?

Answer 1

Okay so it was actually quite simple, you can't use the output of the algorithm straight away, you first need to do this:

layer = self.parameterAsVectorLayer(parameters, self.OUTPUT, context)

Then you can run functions like layer.getFeatures() etc. on that.

Don't know if there are better ways to do what I am trying to do but this works! Anyone who has better ways feel welcome to post a comment.