Latest Contributions by AndrewRudin1

Add the ability to filter a hosted feature layer by a subquery for both definition queries and "select by attribute" workflows. For me, I want to build a hosted view on top of a hosted layer.  The hosted view would limit the results to only the "most recent record" for each ID that exists in a column.  This query requires a date column and the ID column which is not unique.   In a sql database this query would take on the following syntax: For the example the table we are filtering is named myTable. The column with ID's we want to group by is called myGroupID.  The column with dates to use for finding most recent record is named myDate select * from myTable t1 where t1.myDate =  (SELECT max(t2.myDate) from myTable t1 where t2.myGroupID = t1.myGroupID)   Other scenarios are just filtering the whole table to the one with a max value over all: select * from myTable t1 where t1.myDate =  (SELECT max(t2.myDate) from myTable   This came up a lot with Covid dashboard data.  You have a table with one row per day with various Covid numbers.  Then some places you want to filter that master dataset to data from the latest row when sorted by date.  In other places you want to use all the data to see trends.  The operations dashboard widgets provide the extra configuration steps to do this kind of sorting and filtering.  But applying the same kind of filtering to the hosted layer itself via a hosted view is missing.    Now, I assume that on Esri's side this would require the underlying ArcGIS Online data platform to support some kind of SQL structure.  But it wouldn't have to be RDBMS like PostGres or SQL Server.  Azure Synapse allows relatively advanced SQL queries against data lakes that include subqueries and aggregation functions.  Also, I assume that operations like this will add more load on the Esri systems since the filter needs to scan other rows and compare.  But this is handled in RDBMS with indexes.  There is already the concept in ArcGIS Online of tagging one column on a layer to be the "time aware" layer, and the trackaware GeoAnalytics tools in Pro require you to select a "Track ID" column.  If we could define these columns on the AGOL layer, they could be used to selectively build indexes on the backend to optimize these kinds of queries.   ... View more

The new credit usage report is great for showing what users burned in terms of credits over a month or week. However, the report deployed in the Sep 2020 AGOL release does not include credit usage for feature storage, file storage, and tile storage.  These missing categories make up 98% of our org's credit usage, so it doesn't help us much with looking for accounts that are using the most of our credits.   Also, I noticed that the date range at the top of the report is given in UNIX Epoch integer.  A date format that is both human readable and easy to parse with common BI date functions would be better, like ISO standard 8601.  So rather than Aug 1, 2020 UTC appearing as 1596240000, it would be 2020-08-01T00:00:00Z.   ... View more

ArcGIS Pro provides a nice GUI for updating python libraries used by the software.  For folks like me this makes it so much easier to get started with the python features, since I don't have to worry about learning Conda. However, when I hear about a new version of the arcgis python library being available, when I go to the Python Package Manager, it is not shown as an option to update.  I am curious why this flagship library from Esri is not available for update through this window when so many smaller ancillary libraries are available.   I do know that the Python API help has a page explaining how to do the update via the conda terminal:  Install and set up | ArcGIS for Developers .  However, the instructions don't look like they've been updated since 2.2.  Also, it is a little clunky to go through these steps when Pro has a feature to make updates easier.  ... View more

Currently ArcGIS Online allows layers to be non-editable for the public, but if the item is in a "group with update capabilities", the members of that group can add the layer to an AGOL map with full editing control and make edits. However, if these users add these layers to ArcGIS Pro, they cannot edit.  I would like for the group editing ability on the layer be honored in Pro.   I guess the current workaround is to make the hosted layer editable and share that with the group, then make a hosted view that is non-editable, and share that with the public.  It's just extra items for us to maintain. ... View more

I'm joining two tables in ArcGIS Insights (ArcGIS Online) and I know both tables will have orphans after the join.  For my analysis, these orphans need to have the column I'm measuring on converted from Null to zero.  Once I create the joined table, I'm viewing the table and clicking +Field button, with the intent to create a new field that follows this logic: IF [measure_field] is null, THEN Value=0, ELSE Value=[measure_field] In the Insights calculate field window, I couldn't find a way to check if a field's value is null.  Is this possible?  Or is there a better way to cleanup nulls during the join? ... View more

I have some AVL data hosted in a 10.6.1 spatiotemporal big data store.  I'm trying to apply some filters in Pro based on the GPS timestamp but it's not working.  The Pro documentation covers supported functions for the traditional RDMS, shapefile, and file database formats, but doesn't discuss compatibility for the newer ArcGIS Online, Relational Data Store, and Spatiotemporal Data Store options. Specifically I'm trying to do two queries on a single spatiotemporal hosted feature layer and contains historical vehicle location data.  The point layer has a VEHICLE_ID column to store the unique ID for each vehicle, and a GPSTIME field for the time the location was caption. Query #1:  Filter the dataset to only records with a timestamp in the last 24 hours.  I tried a def. query GPSTIME > (CURRENT_TIMESTAMP() - 1), but that returned an expression error. Query #2: Filter the dataset to only the record with the latest timestamp for each vehicle_id.  I believe with a traditional RDMS, this type of query can be handled with a whereclause along the lines of the following:   gpstime = (SELECT max(gpstime) from MYGPSTABLE WHERE vehicle_id = MYGPSTABLE.vehicle_id) ... View more