Feature Layer (archive)

The Feature Layer (archive) feed type in ArcGIS Analytics for IoT loads features exported to the feature layer archive (cold store).

Example

  • A user selects a feature layer that has a data retention export strategy enabled in order to ingest historical data for analysis.

Usage notes

  • When browsing to select a feature layer that has an associated feature layer archive (cold store), you can filter by your content or organization. You can then further filter by date modified, date created, and tags.
  • A feature layer archive only includes features that have been exported and removed from the associated active feature layer. To perform analysis on the entirety of your data, including active features, configure both a Feature Layer and a Feature Layer (archive) source in your big data analytic and use the Merge tool to merge the two data sources together before adding other tools.
  • When data is exported to the feature layer archive (cold store), additional date fields are created that represent the start time (START_TIME) and end time (END_TIME) of the feature. These will appear in the schema of the feature layer archive as DATE and similar. These values are a duplicate of the datetime values in the original field(s) and these additional fields can be dropped if necessary.
  • After configuring source connection properties, see Configure input data to learn how to define the schema and the key properties.

Parameters

ParameterExplanationData Type

Item ID

The item ID of a feature layer that has an associated archive (cold store).

String

Considerations and limitations

  • Only feature layers generated by Analytics for IoT are supported, specifically feature layers for which the data retention policy includes exporting older data to the archive (cold store). For more information, see About data retention.
  • When configuring a data source to load from the feature layer archive, there is currently a limitation with sampling where the sampled features may show incorrect datetime values such as 1858-10-13T08:08:37Z. This issue is limited only to the sampling of values and does not affect successful data ingestion. This issue will be addressed in a future release.