Skip to main content

Hi there,

 

I'm a experienced FME user but have no experience in writing to ArcGIS portal.

I have a hosted feature layer in an ArcGIS portal with three layers. With FME i'am trying to update these layers with new data. I've connected to the portal and imported the feature type definitions from the portal. I've set the layer writer mode to insert and the Truncate first to yes. Tables seems to be truncated and in one layer I see some of the data (670 features instead of 6500). The other layers are empty. There are no errors or warnings in the FME log file.

 

Anybody any idea of what might be wrong?

Hi @arnovananrooij​ 

Before we begin, I am going to make some assumptions, Please check if these are correct.

Your portal is configured with ArcGIS DataStore(based on Postgres).

You are using the same workspace to write new data.

BACKGROUND

Postgres bases ArcGIS Data Store does not support mixed-case attribute names i.e. Attribute --> attribute, ATTRIBUTE --> attribute

During the first run (The first time creating a workspace) all the data is packaged into a zip folder and published to the server, the server unpacks, publishes, and it as Feature Service. During this process, all mixed-case attribute names are changed to lowercase.

Confirm this via login to the ArcGIS Portal WebUI Content --> Your feature service --> Data --> Field

Notice that all field names are in lowercase.

If you open your workspace, Expand the FeatureType and verify if the attribute names are lower case or not.

WORKAROUND

Step 1: Update all the writer Feature Classes

Go to Writer(Drop-down menu) --> Update Writer Feature Types...

Select the Feature layers you want to update

Step 2: Fix incoming data to match schema on the writer

Add BulkAttriubteRenamer transformer before each writer feature type.

Change the transformer properties

Action : Change Case

Case Case Type: lowercase

Please see this Q&A as well.


Hi @rahulsharma​ ,

Like I said, I imported the feature type definitions from the portal (Writers > Import Feature Types), so there can be no mismatch like you're suggesting. It must be something else.

Arno


Hi @arnovananrooij​ 

Can you confirm which build of FME Desktop are you currently using?

Can you try FME Desktop 2021.0 ? We have recently changed some behavior with existing features, where the schema on incoming data (i.e. string length or data type mismatched) are handled better.

It is possible that due to this issue the rest of the batch process is silently failing, you can test this by setting the FeatureType per transaction to the batch of 10 or 1 on the writer to get the exact feature number that is causing the issue.

If you are still unable to resolve the issue, we can escalate this to a support ticket(possible screen share).


Hi @arnovananrooij​ 

Can you confirm which build of FME Desktop are you currently using?

Can you try FME Desktop 2021.0 ? We have recently changed some behavior with existing features, where the schema on incoming data (i.e. string length or data type mismatched) are handled better.

It is possible that due to this issue the rest of the batch process is silently failing, you can test this by setting the FeatureType per transaction to the batch of 10 or 1 on the writer to get the exact feature number that is causing the issue.

If you are still unable to resolve the issue, we can escalate this to a support ticket(possible screen share).

I use FME 2020.2.1.0 (20201130 - Build 20806 - WIN64). Also tried FME(R) 2021.0.1.0 (20210404 - Build 21313 - WIN64), but makes no difference. Still only a few records in one table and two empty tables.


I use FME 2020.2.1.0 (20201130 - Build 20806 - WIN64). Also tried FME(R) 2021.0.1.0 (20210404 - Build 21313 - WIN64), but makes no difference. Still only a few records in one table and two empty tables.

It seems that when I change writer parameter 'features per request' from 1000 to 100 the rsult is much better. At the moment I have all records of two tables. Just one table is still completly emty.


Reply