Use Ibis to return aggregates of every numeric column in a table

Question

I want to use Ibis to return the mean and standard deviation of each floating point numeric column in a table. How do I do that?

Answer 1

Here's a standalone example that uses Ibis to return the mean and standard deviation of each floating point column in the iris flower dataset. This uses Ibis selectors:

import ibis
from ibis import _
import ibis.selectors as s
import ibis.expr.datatypes as dt
import seaborn as sns

iris = ibis.memtable(sns.load_dataset("iris"))

iris.aggregate(
    s.across(
        s.of_type(dt.float()),
        dict(mean=ibis._.mean(), std=ibis._.std())
    )
)

This returns:

┏━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┓
┃ sepal_length_mean ┃ sepal_width_mean ┃ petal_length_mean ┃ petal_width_mean ┃ sepal_length_std ┃ sepal_width_std ┃ petal_length_std ┃ petal_width_std ┃
┡━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━┩
│ float64           │ float64          │ float64           │ float64          │ float64          │ float64         │ float64          │ float64         │
├───────────────────┼──────────────────┼───────────────────┼──────────────────┼──────────────────┼─────────────────┼──────────────────┼─────────────────┤
│          5.843333 │         3.057333 │             3.758 │         1.199333 │         0.828066 │        0.435866 │         1.765298 │        0.762238 │
└───────────────────┴──────────────────┴───────────────────┴──────────────────┴──────────────────┴─────────────────┴──────────────────┴─────────────────┘

You can replace s.of_type(dt.float()) with s.numeric() to apply this to all numeric columns including integers. Or you can use other members of dt to refer to other column types.