featuretools.EntitySet.add_dataframe

EntitySet.add_dataframe(dataframe, dataframe_name=None, index=None, logical_types=None, semantic_tags=None, make_index=False, time_index=None, secondary_time_index=None, already_sorted=False)[source]

Add a DataFrame to the EntitySet with Woodwork typing information.

Parameters
  • dataframe (pandas.DataFrame) – Dataframe containing the data.

  • dataframe_name (str, optional) – Unique name to associate with this dataframe. Must be provided if Woodwork is not initialized on the input DataFrame.

  • index (str, optional) – Name of the column used to index the dataframe. Must be unique. If None, take the first column.

  • logical_types (dict[str -> Woodwork.LogicalTypes/str, optional]) – Keys are column names and values are logical types. Will be inferred if not specified.

  • semantic_tags (dict[str -> str/set], optional) – Keys are column names and values are semantic tags.

  • make_index (bool, optional) – If True, assume index does not exist as a column in dataframe, and create a new column of that name using integers. Otherwise, assume index exists.

  • time_index (str, optional) – Name of the column containing time data. Type must be numeric or datetime in nature.

  • secondary_time_index (dict[str -> list[str]]) – Name of column containing time data to be used as a secondary time index mapped to a list of the columns in the dataframe associated with that secondary time index.

  • already_sorted (bool, optional) – If True, assumes that input dataframe is already sorted by time. Defaults to False.

Notes

Will infer logical types from the data.

Example

In [1]: import featuretools as ft

In [2]: import pandas as pd

In [3]: transactions_df = pd.DataFrame({"id": [1, 2, 3, 4, 5, 6],
   ...:                                 "session_id": [1, 2, 1, 3, 4, 5],
   ...:                                 "amount": [100.40, 20.63, 33.32, 13.12, 67.22, 1.00],
   ...:                                 "transaction_time": pd.date_range(start="10:00", periods=6, freq="10s"),
   ...:                                 "fraud": [True, False, True, False, True, True]})
   ...: 

In [4]: es = ft.EntitySet("example")

In [5]: es.add_dataframe(dataframe_name="transactions",
   ...:                  index="id",
   ...:                  time_index="transaction_time",
   ...:                  dataframe=transactions_df)
   ...: 
Out[5]: 
Entityset: example
  DataFrames:
    transactions [Rows: 6, Columns: 5]
  Relationships:
    No relationships

In [6]: es["transactions"]
Out[6]: 
   id  session_id  amount    transaction_time  fraud
1   1           1  100.40 2022-06-10 10:00:00   True
2   2           2   20.63 2022-06-10 10:00:10  False
3   3           1   33.32 2022-06-10 10:00:20   True
4   4           3   13.12 2022-06-10 10:00:30  False
5   5           4   67.22 2022-06-10 10:00:40   True
6   6           5    1.00 2022-06-10 10:00:50   True