eSearch Widget slow return, looking for ways to speed things up.

993
3
11-17-2020 09:13 AM
James_Armstrong
Occasional Contributor

I am using Robert's eSearch widget in an application created through WAB-DEV.  As usual, his widgets are pretty slick.  

in this instance though, I am running a search against 2,528,552 parcels (from a feature service (state wide parcel set)).  For this particular search, only the 3996 parcels within one town are needed.  There is a definition filter at the webmap level so that only the parcels within the Town are shown.  Similar, in the eSearch widget configuration,  there is a definition expression that (i am assuming) limits the search only the parcels with the Town's ID code. 

Town_ID = 277

The thing is, in using the eSearch, the search for either the Parcel ID or the Owner Name takes along time. Many times to what I am thinking is a time out failure.  Does anyone have suggestions how this search would resolve quicker?  

0 Kudos
3 Replies
RobertScheitlin__GISP
MVP Emeritus

James,

 

  Have you tried this search in the browser just using the rest end point to see if the time needed to perform the search is just due to the speed of your service?

0 Kudos
James_Armstrong
Occasional Contributor

Robert,   

Ok,  I realize I am being a bit dense here.  Im not sure what you mean by:

  Have you tried this search in the browser just using the rest end point to see if the time needed to perform the search is just due to the speed of your service?

Would you expand on you previous responds?  I have included the urls associated with the data set.

0 Kudos
RobertScheitlin__GISP
MVP Emeritus

James,

 

   So what happen in the background of the eSearch is a query task is executed on the REST service. You can manually query a service by clicking on the query operation link at the bottom of your REST service page.

When I execute this query

https://services1.arcgis.com/hGdibHYSPO59RG1h/ArcGIS/rest/services/L3_TAXPAR_POLY_ASSESS_gdb/Feature...

 

Which is a simple query using Town_ID = 277 AND OWNER1 LIKE '%smith%' it takes approximately 99 seconds to execute. 2,528,552 records is just a very large dataset to query.

One way you could speed this up is to have a nightly (or some scheduled) routine to make a separate service for the county of interest.

0 Kudos