rollup() | DataFrame Manipulation | Teradata Package for Python - rollup() Function - Teradata Package for Python

Use the rollup() function to create a multidimensional rollup for a DataFrame using specified columns, and there by running aggregates on it produce the aggregations on different dimensions.

Required Parameter

columns

Specifies the names of input teradataml DataFrame columns.

Optional Parameter

include_grouping_columns

Specifies whether to include aggregations on the grouping columns or not. When set to True, the resultant DataFrame will have the aggregations on the columns mentioned in "columns_expr". Otherwise, resultant DataFrame will not have aggregations on the columns mentioned in "columns_expr".

Default value: False

Example setup

In this example, "admission_train" dataset is used.

>>> from teradataml import *

>>> load_example_data("dataframe", "admissions_train")

>>> df = DataFrame("admissions_train")

Example 1: Analyzes the data by grouping into masters and stats dimensions

>>> df1 = df.rollup(["masters", "stats"]).sum()

>>> df1
  masters     stats  sum_id  sum_gpa  sum_admitted
0      no      None     343    63.96            16
1     yes      None     477    77.71            10
2    None      None     820   141.67            26
3      no    Novice     146    25.41             6
4      no  Beginner       8     3.60             1
5     yes    Novice      98    13.74             1
6     yes  Beginner      13    14.71             2
7     yes  Advanced     366    49.26             7
8      no  Advanced     189    34.95             9

Example 2: Find the average of all valid columns by grouping the DataFrame with columns 'masters' and 'admitted'

Include grouping columns in aggregate function 'avg'.

>>> df1 = df.rollup(["masters", "admitted"], include_grouping_columns=True).avg()
>>> df1
    masters admitted    avg_id  avg_gpa avg_admitted
0        no      NaN 19.055556 3.553333     0.888889
1       yes      NaN 21.681818 3.532273     0.454545
2      None      NaN 20.500000 3.541750     0.650000
3       yes      0.0 24.083333 3.613333     0.000000
4        no      1.0 18.875000 3.595000     1.000000
5       yes      1.0 18.800000 3.435000     1.000000
6        no      0.0 20.500000 3.220000     0.000000
>>>

Example 3: Find the average of all valid columns by grouping the DataFrame with columns 'masters' and 'admitted'

Do not include grouping columns in aggregate function 'avg'.

>>> df1 = df.rollup(["masters", "admitted"], include_grouping_columns=False).avg()
>>> df1
    masters admitted    avg_id  avg_gpa
0        no      NaN 19.055556 3.553333
1       yes      NaN 21.681818 3.532273
2      None      NaN 20.500000 3.541750    
3       yes      0.0 24.083333 3.613333
4        no      1.0 18.875000 3.595000
5       yes      1.0 18.800000 3.435000
6        no      0.0 20.500000 3.220000
>>>