Contents

Slicing and selecting using .iloc[]

Watch it

See the accompanied youtube video at the link here.

Up until point, we have been manipulating our dataframe with column and row labels using .loc[].

Slicing can also be done by the location position of each row with .iloc[].

.iloc[] is very similar to .loc[], however, the “i” in iloc refers to the index integer position.

Slicing Dataframe

We are going to return to our cereal dataset and take a look at the first 10 rows.

cereal = pd.read_csv('cereal.csv')
cereal.head(10)
name mfr type calories protein fat sodium ... sugars potass vitamins shelf weight cups rating
0 100% Bran N Cold 70 4 1 130 ... 6 280 25 3 1.00 0.33 68.402973
1 100% Natural Bran Q Cold 120 3 5 15 ... 8 135 0 3 1.00 1.00 33.983679
2 All-Bran K Cold 70 4 1 260 ... 5 320 25 3 1.00 0.33 59.425505
3 All-Bran with Extra Fiber K Cold 50 4 0 140 ... 0 330 25 3 1.00 0.50 93.704912
4 Almond Delight R Cold 110 2 2 200 ... 8 1 25 3 1.00 0.75 34.384843
5 Apple Cinnamon Cheerios G Cold 110 2 2 180 ... 10 70 25 1 1.00 0.75 29.509541
6 Apple Jacks K Cold 110 2 0 125 ... 14 30 25 2 1.00 1.00 33.174094
7 Basic 4 G Cold 130 3 2 210 ... 8 100 25 3 1.33 0.75 37.038562
8 Bran Chex R Cold 90 2 1 200 ... 6 125 25 1 1.00 0.67 49.120253
9 Bran Flakes P Cold 90 3 0 210 ... 5 190 25 3 1.00 0.67 53.313813

10 rows × 16 columns

Let’s say we want the rows from All-Bran to Apple Cinnamon Cheerios, but we want to slice based on their position instead of their label.

Using Python’s counting method of starting at zero, we conclude that All-Bran to be at position to 2.

We get Apple Cinnamon Cheerios position to be 5 in the same way.

We are lucky with this dataframe because our index labels match the position of the rows, and this makes things a little bit simpler.

We can use the same coding structure we learned with .loc[] but this time using row positions instead of labels with .iloc[].

cereal.loc[2:5]
name mfr type calories protein fat sodium ... sugars potass vitamins shelf weight cups rating
2 All-Bran K Cold 70 4 1 260 ... 5 320 25 3 1.0 0.33 59.425505
3 All-Bran with Extra Fiber K Cold 50 4 0 140 ... 0 330 25 3 1.0 0.50 93.704912
4 Almond Delight R Cold 110 2 2 200 ... 8 1 25 3 1.0 0.75 34.384843
5 Apple Cinnamon Cheerios G Cold 110 2 2 180 ... 10 70 25 1 1.0 0.75 29.509541

4 rows × 16 columns

cereal.iloc[2:5]
name mfr type calories protein fat sodium ... sugars potass vitamins shelf weight cups rating
2 All-Bran K Cold 70 4 1 260 ... 5 320 25 3 1.0 0.33 59.425505
3 All-Bran with Extra Fiber K Cold 50 4 0 140 ... 0 330 25 3 1.0 0.50 93.704912
4 Almond Delight R Cold 110 2 2 200 ... 8 1 25 3 1.0 0.75 34.384843

3 rows × 16 columns

But wait! Something is missing here!

Why doesn’t Apple Cinnamon Cheerios appear in the dataframe?

The reason for this is that when we use slicing with indices, it will take all the indices including the lower bound but EXCLUDING the upper bound.

If we want to include Apple Cinnamon Cheerios we would have to go 1 index position further.

cereal.iloc[2:6]
name mfr type calories protein fat sodium ... sugars potass vitamins shelf weight cups rating
2 All-Bran K Cold 70 4 1 260 ... 5 320 25 3 1.0 0.33 59.425505
3 All-Bran with Extra Fiber K Cold 50 4 0 140 ... 0 330 25 3 1.0 0.50 93.704912
4 Almond Delight R Cold 110 2 2 200 ... 8 1 25 3 1.0 0.75 34.384843
5 Apple Cinnamon Cheerios G Cold 110 2 2 180 ... 10 70 25 1 1.0 0.75 29.509541

4 rows × 16 columns

If we think about this a bit it actually make some sense. Think about the calculation 6 - 2 = 4 . We get 4 items remaining which is the amount of cereals we want in our in new dataframe.

The same concept can be applied to the columns of the dataframe.

Let’s say we want all the rows but we only want the columns starting at name and ending (and including) at column fat.
Using the logic we learned in the last section, we would use the following code.

We would need to specify all rows using : as we did when we used .loc[].

The column name is at index position 0 (we do not include the index label as a column) and fat is at index position 5.

Since we want to include the 5th column we need to use the 6th position to make sure we get all the columns BEFORE the upper bound.

cereal.iloc[:, 0:6]
name mfr type calories protein fat
0 100% Bran N Cold 70 4 1
1 100% Natural Bran Q Cold 120 3 5
2 All-Bran K Cold 70 4 1
3 All-Bran with Extra Fiber K Cold 50 4 0
4 Almond Delight R Cold 110 2 2
... ... ... ... ... ... ...
72 Triples G Cold 110 2 1
73 Trix G Cold 110 1 1
74 Wheat Chex R Cold 100 3 1
75 Wheaties G Cold 100 3 1
76 Wheaties Honey Gold G Cold 110 2 1

77 rows × 6 columns

Let’s say we want the rows All-Bran to Apple Cinnamon Cheerios and name to fat.

For rows, the lower bound All-Bran is located at position 2 and the upper bound Apple Cinnamon Cheerios is located at position 5.

Now the column’s lower bound name is located at position 0 and the upper bound fat is located at position 5.

The same would apply if we only wanted certain rows with certain columns.

Both of our upper bound have been compensated with an added 1 to make sure they are included in the new dataframe.

So the code we have to use to do this is the following:

cereal.iloc[2:6, 0:6]
name mfr type calories protein fat
2 All-Bran K Cold 70 4 1
3 All-Bran with Extra Fiber K Cold 50 4 0
4 Almond Delight R Cold 110 2 2
5 Apple Cinnamon Cheerios G Cold 110 2 2

There are multiple different way of writing code when you are selecting items from the beginning or end of your data.

Perhaps you only want the first 3 rows of your data.

We can use .head(3) or we can use .iloc[].

cereal.head(3)
name mfr type calories protein fat sodium ... sugars potass vitamins shelf weight cups rating
0 100% Bran N Cold 70 4 1 130 ... 6 280 25 3 1.0 0.33 68.402973
1 100% Natural Bran Q Cold 120 3 5 15 ... 8 135 0 3 1.0 1.00 33.983679
2 All-Bran K Cold 70 4 1 260 ... 5 320 25 3 1.0 0.33 59.425505

3 rows × 16 columns

cereal.iloc[0:3]
name mfr type calories protein fat sodium ... sugars potass vitamins shelf weight cups rating
0 100% Bran N Cold 70 4 1 130 ... 6 280 25 3 1.0 0.33 68.402973
1 100% Natural Bran Q Cold 120 3 5 15 ... 8 135 0 3 1.0 1.00 33.983679
2 All-Bran K Cold 70 4 1 260 ... 5 320 25 3 1.0 0.33 59.425505

3 rows × 16 columns

Since we are indicating the beginning of the dataframe, we can omit the upper bound 0 just like we did when we learned slicing with .loc[].

cereal.iloc[:3]
name mfr type calories protein fat sodium ... sugars potass vitamins shelf weight cups rating
0 100% Bran N Cold 70 4 1 130 ... 6 280 25 3 1.0 0.33 68.402973
1 100% Natural Bran Q Cold 120 3 5 15 ... 8 135 0 3 1.0 1.00 33.983679
2 All-Bran K Cold 70 4 1 260 ... 5 320 25 3 1.0 0.33 59.425505

3 rows × 16 columns

The same logic can be applied for end of a dataframe. This time we want the last 3 rows.

Since this dataframe has 76 rows we could use our lower bound as 74 and upper bound as 77 (76 +1).

cereal.iloc[74:77]
name mfr type calories protein fat sodium ... sugars potass vitamins shelf weight cups rating
74 Wheat Chex R Cold 100 3 1 230 ... 3 115 25 1 1.0 0.67 49.787445
75 Wheaties G Cold 100 3 1 200 ... 3 110 25 1 1.0 1.00 51.592193
76 Wheaties Honey Gold G Cold 110 2 1 200 ... 8 60 25 1 1.0 0.75 36.187559

3 rows × 16 columns

A better and easier way to write this, where you don’t even need to know the number of rows in the dataframe would be to specify you are counting your rows from the bottom with a negative in front of the number of rows you want.

This example takes the last 3 rows of the bottom of the dataframe.

Since we are collecting data to the end of the dataframe, we do not need to include the ending row index number.

cereal.iloc[-3:]
name mfr type calories protein fat sodium ... sugars potass vitamins shelf weight cups rating
74 Wheat Chex R Cold 100 3 1 230 ... 3 115 25 1 1.0 0.67 49.787445
75 Wheaties G Cold 100 3 1 200 ... 3 110 25 1 1.0 1.00 51.592193
76 Wheaties Honey Gold G Cold 110 2 1 200 ... 8 60 25 1 1.0 0.75 36.187559

3 rows × 16 columns

Selecting with .iloc[]

Selecting using .iloc[] is done identically to .loc[], however, the items within each set of square brackets MUST be integers, and not in quotation marks.

cereal.head(10)
name mfr type calories protein fat sodium ... sugars potass vitamins shelf weight cups rating
0 100% Bran N Cold 70 4 1 130 ... 6 280 25 3 1.00 0.33 68.402973
1 100% Natural Bran Q Cold 120 3 5 15 ... 8 135 0 3 1.00 1.00 33.983679
2 All-Bran K Cold 70 4 1 260 ... 5 320 25 3 1.00 0.33 59.425505
3 All-Bran with Extra Fiber K Cold 50 4 0 140 ... 0 330 25 3 1.00 0.50 93.704912
4 Almond Delight R Cold 110 2 2 200 ... 8 1 25 3 1.00 0.75 34.384843
5 Apple Cinnamon Cheerios G Cold 110 2 2 180 ... 10 70 25 1 1.00 0.75 29.509541
6 Apple Jacks K Cold 110 2 0 125 ... 14 30 25 2 1.00 1.00 33.174094
7 Basic 4 G Cold 130 3 2 210 ... 8 100 25 3 1.33 0.75 37.038562
8 Bran Chex R Cold 90 2 1 200 ... 6 125 25 1 1.00 0.67 49.120253
9 Bran Flakes P Cold 90 3 0 210 ... 5 190 25 3 1.00 0.67 53.313813

10 rows × 16 columns

Let’s say we want the rows Almond Delight, Basic 4 and Apple Jacks with the columns name, calories, fat and type and in that specific order.

Row

Row Position

Almond Delight

Position 4

Basic 4

Position 7

Apple Jacks

Position 6

Columns

Column Position

name

Position 0

calories

Position 3

fat

Position 5

type

Position 2

Almond Delight takes row position 4, Basic 4 takes row position 7 and Apple Jacks is located at position 6. The desired columns name, calories, fat, and type take column index positions 0, 5, 3, and 2 respectively.

Now let’s put those positions into square backing within df.iloc[].

cereal.iloc[[4, 7, 6], [0, 3, 5, 2]]
name calories fat type
4 Almond Delight 110 2 Cold
7 Basic 4 130 2 Cold
6 Apple Jacks 110 0 Cold

Let’s apply what we learned!

Here is our fruit_salad data again:

           name    colour    location    seed   shape  sweetness   water-content  weight
0         apple       red     canada    True   round     True          84         100
1        banana    yellow     mexico   False    long     True          75         120
2    cantaloupe    orange      spain    True   round     True          90        1360
3  dragon-fruit   magenta      china    True   round    False          96         600
4    elderberry    purple    austria   False   round     True          80           5
5           fig    purple     turkey   False    oval    False          78          40
6         guava     green     mexico    True    oval     True          83         450
7   huckleberry      blue     canada    True   round     True          73           5
8          kiwi     brown      china    True   round     True          80          76
9         lemon    yellow     mexico   False    oval    False          83          65

1. If I wanted the rows elderberry to kiwi and only columns seeds, shape, sweetness and water-content, what would my code look like if I was using index positions?
a) fruit_salad.iloc[4:9, 3:]
b) fruit_salad.iloc[4:8, 3:7]
c) fruit_salad.iloc[4:9, 3:7]
d) fruit_salad.iloc[5:10, 4:8]

2. If I wanted the rows lemon and cantaloupe but only the columns colour, weight and seeds in that order, what would my code look like if I was using index position?
a) fruit_salad.iloc[[lemon, cantaloupe], [colour, weight, seeds]]
b) fruit_salad.iloc[[10, 3], [2, 8, 4]]
c) fruit_salad.iloc[[9, 2], [1, 7, 3]]
d) fruit_salad[[9, 2], [1, 7, 3]]

Solutions!

  1. c) fruit_salad.iloc[4:9, 3:7]

  2. c) fruit_salad.iloc[[9, 2], [1, 7, 3]]