ArcGIS dashboards Data Expression including Rank
Sorting and looping through would work with a single field, but even that can take a while to execute. Adding multiple fields, well, that will be a really long-running expression.
There may be a better way to approach this, but it will still result in tons of queries to the server. You could try using a custom function to "memorize" the FeatureSet first, which will speed things up a lot. (https://www.jcarlson.page/posts/dashboards-memorize/)
To get the rank of each field in a single loop, we just need some filtering. You can do a filter and see how many features are greater/less than the current feature, and that's basically the rank. Maybe add 1 to it? The smallest value in a field will have no features less than it, so the filtered count + 1 = 1, that's it's rank.
Since we're doing this repeatedly for multiple fields we could use another custom function for the rank.
Here's an example, just swap in your values as needed:
function Memorize(fs) {
var temp_dict = {
fields: Schema(fs)['fields'],
geometryType: '',
features: []
}
for (var f in fs) {
var attrs = {}
for (var attr in f) {
attrs[attr] = Iif(TypeOf(f[attr]) == 'Date', Number(f[attr]), f[attr])
}
Push(
temp_dict['features'],
{attributes: attrs}
)
}
return FeatureSet(Text(temp_dict))
}
function getRank(feat, field, set){
var filt_val = feat[field]
var filt_set = Filter(set, `${field} < @filt_val`)
return Count(filt_set)
}
var fs = Memorize(Filter(FeatureSetByPortalItem(
Portal("https://www.arcgis.com"),
"6200db0b80de4341ae8ee2b62d606e67",
0,
["*"],
false
), 'OBJECTID < 100'))
// ranks to add to the output
var rank_fields = [
{name: 'objectid_rank', type:'esriFieldTypeInteger'},
{name: 'eaveheight_rank', type:'esriFieldTypeInteger'},
{name: 'area_rank', type:'esriFieldTypeInteger'}
]
var out_dict = {
fields: Splice(Schema(fs)['fields'], rank_fields),
geometryType: '',
features: []
}
var full_count = Count(fs)
for (var f in fs) {
// convert feature to dict
var feat_dict = FromJSON(Text(f))
// calculate ranks
feat_dict['attributes']['objectid_rank'] = getRank(f, 'OBJECTID', fs)
feat_dict['attributes']['eaveheight_rank'] = getRank(f, 'EAVEHEIGHT', fs)
feat_dict['attributes']['area_rank'] = getRank(f, 'SHAPE__Area', fs)
// convert back to feature, push into output
Push(out_dict['features'], Feature(feat_dict))
}
return FeatureSet(Text(out_dict))
The error is in the getRank function. The SQL statement should be a backtick string, sorry!
`${field} < @filt_val`
To answer your questions:
The entire "Memorize" function is just taking your FeatureSet and pushing it into a new FeatureSet that is held in your browser's memory. That way future requests for information (like the Filter and Count functions) work with local data. Normally, your expression would have to wait for queries to be sent to the feature service each time.
On "GetRank", it doesn't identify the smallest feature. One of the parameters is "feat", and the function can be called on any individual feature. Since we're just defining the function, it's not being used in any particular order.
To explain it again, imagine you have 8 values. If you simply take each value and count the number of features that are smaller than it, that is the rank. You don't actually need to go through each feature counting by ones.
Full count isn't being used, it was an earlier idea I forgot to remove. You can delete it!
