Available in real-time and big data analytics.
The Filter by Geometry tool 
filters features by a spatial relationship with other features in the same feed or those of another feed or dataset. Features that the spatial relationship evaluates as true are retained.
Examples
The following are example uses of the Filter by Geometry tool:
- A real-time analytic uses the Filter by Geometry tool to retain only the incoming records from a feed in which the feature geometry intersects a polygon dataset of port areas.
- A big data or real-time analytic uses the Filter by Geometry tool to retain only features in which the feature geometry enters a polygon dataset representing service areas.
- Determine when shipping vessels from one feed are inside a moving severe weather cell from another feed.
Usage notes
Keep the following in mind when working with the Filter by Geometry tool:
- The tool can compare the spatial relationship of features to point, line, or polygon features.
- The tool evaluates a defined spatial relationship between the target features and a specified join dataset.
- If a Spatial relationship value of Enter or Exit is chosen, checking the First observation can trigger enter or First observation can trigger exit check box indicates that the first target feature received satisfies the filter condition if it is inside any join feature (in the case of Enter) or outside all join features (in the case of Exit), despite having no previous target feature to compare to the active target feature's location. The default is unchecked.
- If the Spatial relationship parameter value is set to Enter or Exit in a real-time analytic, the tool operates in a stateful manner. This enables it to compare sequential observations. For example, a change in state is detected as a current condition different from that of the previous observation. With other spatial relationships in real-time analytics, the tool does not need to compare observations with previous observations, so it can run in a stateless manner, which is less resource intensive.
- When the tool is operating in a stateful manner, (Enter or Exit is chosen for the Spatial relationship value), the Target time window parameter is required. This parameter should be set to the longest anticipated interval between observations for any given track at a minimum. Observations older than this duration are deleted from memory to manage resources.
- It is recommended that you set this value in excess, as a value that is too short would result in deleting records from the feature store before new observations arrive. This means the historical information is deleted from the feature store and the next feature is treated as the initial observation. In other words, observations would never enter or exit the Filter by Geometry tool because the analytic would have no knowledge of the previous observation.
- When operating in a stateful manner, the Filter by Geometry tool maintains a state store of the active observation for each track ID. When a new observation is received, it is compared to the active observation of the same track ID, if present. If the new observation has a newer time stamp than the active observation in the state store, the active observation is replaced. If it has an identical time stamp to the current observation for the same track ID, the tool does not distinguish between the two observations. This means the tool does not replace the active observation in the state store with the new observation, the new observation is not added to the state store, and future observations for this track ID are compared to the active observation that remained in the state store until a new observation arrives with a later time stamp. A warning is logged by the tool when this occurs.
- In dynamic geofencing, the Join time window parameter must be set. If the join feed does not have a field tagged as END_TIME and the last known observation for a join feature is older than this window, it is deleted from the tool's memory and is not included in the analysis. If the join feed has a field tagged as END_TIME, the feature ages out of the geofence store according to the value in the field tagged as END_TIME or at the close of the join time window, whichever comes first. Restarting a real-time analytic does not cause features to expire from the geofence store.
Parameters
The following are parameters for the Filter by Geometry tool:
| Parameter | Description | Data type |
|---|---|---|
Input Layer | Specifies the layer that is filtered. | Features |
Join Layer | Specifies the layer to process a spatial relationship. | Features |
Spatial relationship | Specifies the criteria used to spatially filter the target features. Spatial relationships include the following:
Note:A target feature must satisfy its spatial relationship to all join features when Disjoint or Exit has been specified and there is more than one join feature. For any other spatial relationship and where there is more than one join feature, a target feature can satisfy its spatial relationship to any individual join feature. | String |
Target time window | When operating in a stateful manner, this parameter should be set to the longest anticipated interval between observations for any given track at a minimum. Observations older than this duration are deleted from memory to manage resources. | String |
Join time window | The time window for the join feed (dynamic geofencing). If the last known observation for a join feature is older than the time window specified, it is deleted and is not included in the analysis. If the join feed has a field marked as END_TIME, this parameter is optional. | String |
Output layer
The Filter by Geometry tool does not alter incoming features; the output layer contains the same fields and attribute values as the incoming features.