Far from ideal and far from accurate, I currently use the below calculation. Although it does not get the actual nearest point, it does provide an indication of the nearest by limiting the search distance. For our usecase this is sufficient to give an indication of the nearest street name. This is purely to provide an indication, not important for the actual data.
pulldata("@layer", "getValueAt", "attributes.<ATTRIBUTE>", "https://<URL>/arcgis/rest/services/<SERVICE>/FeatureServer/<LAYERID>?distance=25&units=esriSRUnit_Meter", ${GEOMETRY_NAME})
What I use as a better solution is in a web map use an Arcade expression to find the actual nearest feature. Admittedly, I have the advantage we only use Survey123 to get reports which ArcGIS Online users then edit using a web map. So I am lucky I can use Arcade expressions in a Form. That Arcade expression is adapted from From the Smart Editor to Smart Forms.
var wegvak = FeatureSetByName($map, "<POLYGONLAYER_STREETNAMES>")
var bufferedLocation = Buffer($feature, 500, 'meters')
var candidateWegvakken = Intersects(wegvak, bufferedLocation)
var featuresWithDistances = []
for (var f in candidateWegvakken) {
Push(featuresWithDistances,
{
'distance': Distance($feature, f, 'feet'),
'feature': f
}
)
}
// Sort the candidates by distance using a custom function
function sortByDistance(a, b) {
return a['distance'] - b['distance']
}
var sorted = Sort(featuresWithDistances, sortByDistance)
// Get the closest candidate
var closestFeatureWithDistance = First(sorted)
// If there was no candidate, return null
if (IsEmpty(closestFeatureWithDistance)) { return null }
// Return the candidate name attribute value
return `${closestFeatureWithDistance['feature']['<STREET_NAME>']}`
Frankly, it's great and all we can use AI in Survey123, but we still cannot properly use Arcade expressions and use other coordinate systems in Survey123. I think the priorities are mixed up here..
