Defining structured dtypes in NumPy - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is a structured dtype in NumPy?

A structured dtype in NumPy is a way to define a data type that contains multiple named fields, each with its own type. It allows you to store complex records like a table with columns of different types.

beginner

How do you define a structured dtype with fields 'name' (string) and 'age' (integer)?

You define it using a list of tuples: [('name', 'U10'), ('age', 'i4')]. 'U10' means a Unicode string of max length 10, and 'i4' means a 4-byte integer.

beginner

Why use structured dtypes instead of regular arrays?

Structured dtypes let you store different types of data together in one array, like a spreadsheet row. Regular arrays hold only one data type, so structured dtypes are useful for mixed data.

beginner

What does this dtype definition mean? dtype = [('x', 'f4'), ('y', 'f4')]

It defines a structured dtype with two fields: 'x' and 'y', both 4-byte floating point numbers. This can represent 2D points with float coordinates.

beginner

How do you access the 'age' field from a structured NumPy array named 'data'?

You access it by using data['age']. This returns an array of all the 'age' values from each record.

What is the correct way to define a structured dtype with fields 'height' (float) and 'weight' (float)?

A [('height', 'int'), ('weight', 'int')]

B [('height', 'f8'), ('weight', 'f8')]

C ('height', 'weight')

D ['height', 'weight']

Option B correctly defines a structured dtype with two float fields using a list of tuples.

If you have a structured array 'arr' with a field 'score', how do you get all scores?

A arr.get('score')

B arr.score

C arr['score']

D arr.score()

You access fields in structured arrays using square brackets and the field name as a string.

What does 'U10' mean in a structured dtype field?

A Unicode string of max length 10

B 10-byte unsigned integer

C Unsigned integer of 10 bits

D 10-character ASCII string

'U10' means a Unicode string with a maximum length of 10 characters.

Which of these is NOT a benefit of structured dtypes?

A Store multiple data types in one array

B Access fields by name

C Represent complex records

D Automatically sort data

Structured dtypes do not automatically sort data; sorting is a separate operation.

How would you define a structured dtype for a record with a 2D point (x, y) as floats and a label as a string of max length 5?

A [('x', 'f4'), ('y', 'f4'), ('label', 'U5')]

B [('x', 'i4'), ('y', 'i4'), ('label', 'U10')]

C [('point', 'f8'), ('label', 'U5')]

D [('x', 'f8'), ('y', 'f8'), ('label', 'int')]

Option A correctly defines two float fields for x and y and a Unicode string label of length 5.

Explain how to create and use a structured dtype in NumPy to store records with different data types.

Think of a table with columns of different types stored in one array.

Describe the advantages of using structured dtypes over regular NumPy arrays.

Consider why you might want to keep related data together with different types.