Máscaras booleanas para seleção

Write a comparison and NumPy applies it to every element at once. The result is a boolean array of True and False values.

This array of True and False is a mask. Each position marks whether that element passed your condition.

Put the mask in square brackets to keep only the True positions. NumPy returns just the elements that matched.

This replaces a manual loop with one expression. Boolean filtering is faster to write and far faster to run on big arrays.

Join masks with & for and and the pipe for or. Each condition must sit inside its own parentheses.

Without parentheses around each condition, Python applies operators in the wrong order and raises an error. Wrap every comparison.

The ~ operator inverts a mask, turning True into False. Use it to select everything that did not match your condition.

Booleans act like 1 and 0, so sum on a mask counts how many elements passed. It is a quick way to size a subset.

Use any to ask if at least one element matched, and all to ask if every element did. Both return a single True or False.

A mask can target elements for assignment, not just reading. Here every value above 10 is capped at 10 in place.

From dropping outliers to fixing bad entries, masks are how data scientists select and edit exactly the values they care about.

You want values both above 5 and below 20.

You built boolean masks, filtered and counted with them, combined conditions safely, and even assigned through them. That is selection mastery. ✨