How to Update attributes in an Online Feature Layer using Selection

[ChuckBenton]

I have a large feature layer hosted in ArcGIS Online. I want to select a subset by matching a field attribute to a value, and then update another field to specific value. Ideally I'd prefer not to iterate through any loops as the overall layer is > 150M records. The selection field is indexed and performs quickly, I just can't figure out how to modify/update the other field based on the selection. I've seen arcpy solutions that work with shapefiles and fgdbs, but nothing for Online.

Thanks in advance for any advice or pointers! 

[DavinWalker2 · Accepted solution]

Hi Chuck,

I think  FeatureLayer.calculate is what you are looking for. 

[ChuckBenton]

I'm sure this is a simple issue, but it's stumped me.

The following line executes without issue:

fs_layer.calculate(where="FIPS=18039", calc_expression={"field":"VERSION", "value": "07/04/2021"})

This fails:

where_str = '"FIPS=18039"'
calc_str = '"field":"VERSION", "value":"07/10/2021"'
fs_layer.calculate(where=where_str, calc_expression={calc_str})

I've multiple ways to build the parameters to be passed, building the entire argument, varying quote schemes. What am I missing?

[ConradSchaefer__DOIT_]

Chuck,

looks like the difference is a dict versus a set containing a string.

This is a valid dict (key: value) and looks to be an acceptable parameter type for calculate

• {"field":"VERSION", "value": "07/04/2021"}

Simplifying your code your second option is:

• {'      "field":"VERSION",      "value":"07/10/2021"       '} I added spaces for clarity
• notice the entire inner contents are quoted by ''
• See sets https://python-reference.readthedocs.io/en/latest/docs/sets/
• Your code puts a string inside of curly brackets ... calc_expression={calc_str}

[MirHashmi]

@DavinWalker2 Hi Davin, it not does work with me when I am passing my own value to be updated for an attribute. However, it works if I pass the column name  ? 

This does not work:  fs_layer.calculate(where="FIPS=18039", calc_expression={"field":"FIELDNAME1", "value":"newattributevalue"}) // this returns an error PerformEdit() failure. ColumnName='newattributevalue' not found in field name map. (Error Code: 400)

This works: fs_layer.calculate(where="FIPS=18039", calc_expression={"field":"FIELDNAME1", "value":"FIELDNAME2"}) // this one copies the value from FIELDNAME2 to FIELDNAME1. Is this how calculate works?

Thank you

[emedina]

For this kind of thing, I'd recommend familiarizing yourself with Spatially Enabled DataFrames: https://developers.arcgis.com/python/guide/introduction-to-the-spatially-enabled-dataframe/

Your general process will look like this:

1. Create a FeatureLayer object from the layer you want to update
2. Create an SEDF from the FeatureLayer
3. Perform your selections/manipulations in the SEDF. When satisfied with the result, optionally create a version of the dataframe that only includes the field(s) you're updating and the unique identifier used for the update (ObjectID/GlobalID).
4. Convert the result dataframe to a feature set.
5. Execute an update operation with the FeatureLayer object's  "edit_features" method (your update is the feature set you created).

Hope this helps.

[FrancisCorbett]

This worked great - thanks for the workflow

[ChuckBenton]

Thanks for the responses! featurelayer.calculate fit the bill. I'd skipped over it in my searches, because I'm not doing any "calculations". Once found it was an easy implementation. Thanks!!!

Chuck

[ConradSchaefer__DOIT_]

.calculate() proved to be very simple and precise. Didn't know it existed before but definitely valuable for very precise editing of individual field(s).