Select

Pick one or more columns from a wide (rows, columns) value stream by position. The index and the number of rows are unchanged; only the width changes. Usable eagerly and inside a Pipeline.

Feeding a lazy iterator of (value, index) pairs returns a lazy iterator of column-selected events; feeding arrays or (values, index) tuples returns the batch result.

Example

Keep column 1 of a two-column stream.

import numpy as np
from screamer import Select

wide = np.array([[10.0, 11.0],
                 [20.0, 21.0],
                 [30.0, 31.0]])

column, idx = Select(1)(wide)
print(column)

[11. 21. 31.]

Pass a list of positions (for example [0, 2]) to keep several columns in a chosen order.