Filter Paths

Filters entire paths based on computed path-level metrics. Paths that do not match the condition are removed entirely (all their events are dropped). Unlike Filter Events, this operates at the path level.

Usage

# Keep paths with more than 5 events
es.filter_paths({"op": ">", "metric": "length", "value": 5})

# Keep paths that contain a purchase
es.filter_paths({"op": "=", "metric": "has", "metric_args": {"events": "purchase"}, "value": True})

# Combine conditions
es.filter_paths({
    "op": "and",
    "args": [
        {"op": ">", "metric": "length", "value": 3},
        {"op": "<", "metric": "duration", "value": 3600},
    ]
})

Parameters

Parameter	Type	Description
`ast_condition`	`dict`	A nested condition tree. See below.
`path_id_col`	`str \| None`	Override the path ID column.
`event_col`	`str \| None`	Override the event column.

Condition tree

The ast_condition is a nested dict. Leaf nodes compare a metric against a value. Internal nodes combine leaves with logical operators. See Path Metrics for the full list of available metrics and their arguments.

Logical operators

Op	Structure
`and`	`{"op": "and", "args": [...]}`
`or`	`{"op": "or", "args": [...]}`
`not`	`{"op": "not", "args": [...]}`

Comparison operators

=, !=, >, <, >=, <=, in

Available metrics

Metric	metric_args	Description
`length`	—	Number of events in the path.
`duration`	—	Total path duration in seconds.
`has`	`events`: str or list	Whether the path contains the event(s).
`event_count`	`event`: str or list	Number of times an event occurred.
`time_between`	`event_from`, `event_to`	Seconds between first occurrences of two events.
`active_days`	—	Number of distinct calendar days with activity.
`matches`	`pattern`: str	Whether the path matches a sequence pattern.