verticapy.vDataFrame.sessionize#

vDataFrame.sessionize(ts: str, by: str | list[str] | None = None, session_threshold: str = '30 minutes', name: str = 'session_id') → vDataFrame#

Adds a new vDataColumn to the vDataFrame that corresponds to sessions (user activity during a specific time). A session ends when ts - lag(ts) is greater than a specific threshold.

Parameters#

ts: str: vDataColumn used as timeline. It is used to order the data. It can be a numerical or type date (date, datetime, timestamp…) vDataColumn.
by: SQLColumns, optional: vDataColumn used in the partition.
session_threshold: str, optional: This parameter is the threshold that determines the end of the session. For example, if it is set to ‘10 minutes’, the session ends after 10 minutes of inactivity.
name: str, optional: The session name.

Returns#

vDataFrame: self

Examples#

Let’s begin by importing VerticaPy.

import verticapy as vp

Hint

By assigning an alias to verticapy, we mitigate the risk of code collisions with other libraries. This precaution is necessary because verticapy uses commonly known function names like “average” and “median”, which can potentially lead to naming conflicts. The use of an alias ensures that the functions from verticapy are used as intended without interfering with functions from other libraries.

Let us create a vDataFrame with multiple columns:

vdf = vp.vDataFrame(
    {
        "time": [
            "1993-11-03 00:00:00",
            "1993-11-03 00:14:00",
            "1993-11-03 00:07:00",
            "1993-11-03 01:00:00",
            "1993-11-03 01:05:05",
            "1993-11-03 01:15:05",
            "1993-11-03 01:45:01",
        ],
        "val": [0., 1., 2., 4., 5., 5.5, 6., 8],
    },
)

We can make sure the correct data type is assigned:

vdf["time"].astype("datetime")

	📅 time Datetime 100%	123 val Numeric(4) 100%
1	1993-11-03 00:00:00	0.0
2	1993-11-03 00:14:00	1.0
3	1993-11-03 00:07:00	2.0
4	1993-11-03 01:00:00	4.0
5	1993-11-03 01:05:05	5.0
6	1993-11-03 01:15:05	5.5
7	1993-11-03 01:45:01	6.0

Now we can create sessions based on custom time intervals:

vdf.sessionize(
    ts = "time",
    session_threshold = "15 minutes",
)

	📅 time Datetime 100%	...	123 val Numeric(4) 100%	123 session_id Integer 100%
1	1993-11-03 00:00:00	...	0.0	0
2	1993-11-03 00:07:00	...	2.0	0
3	1993-11-03 00:14:00	...	1.0	0
4	1993-11-03 01:00:00	...	4.0	1
5	1993-11-03 01:05:05	...	5.0	1
6	1993-11-03 01:15:05	...	5.5	1
7	1993-11-03 01:45:01	...	6.0	2

Note

This method is particularly useful when dealing with clickstream analytics, enabling the creation of sessions as part of data preparation for machine learning. It helps understand user behaviors and interactions as they navigate through applications, for example.