Inspect DataFrames¶

Captain’s Inspection

Aarla went for a round of inspection through the ship. She always did before departing on a journey. It was a captains responsibility to make sure every room was in order. Also, there should be some tasty fish soup in the fridge! ...

The ships computer has prepared a tabular report already. All you need to do is to see what is in the table.

Fasten your seat belts and load the report stations.csv into a DataFrame and check the following commands.

What is a DataFrame?¶

The class polars.DataFrame is the central data structure in polars. You can think of it as a table of data, with columns (each of a specific type) and rows. Unlike pandas, Polars does not have an implicit index; all data is stored in columns, and if you want an “index,” you create it explicitly. Each column is a polars.Series. Polars is designed for high-performance and parallel computation, making it especially fast on large datasets.

Inspect a few rows¶

You might want to see a few rows. In Jupyter, you can do that by typing the variable name in a cell. In a Python script, you could print(df). But you could use the more explicit methods head() and tail():

df.head(3)
df.tail(3)

Number of rows and columns¶

The shape of a DataFrame is alway a tuple of two integers (rows, columns).

df.shape

Data types¶

When you load data from a CSV file, polars automatically infers data types. Sometimes a single wrong value converts a numerical column to strings.

df.dtypes

Generic overview¶

Polars prints the table by default with data types, so often df is already enough for a quick overview:

df

You can check data types and the number of entries for each column:

df.glimpse()

Unique values¶

With categorical columns, you might want to know, what are the most frequent values or what different values occur. This also helps you to identify some data errors.

df['column_name'].value_counts()

If you are not interested in the count, check the unique values:

df['column_name'].unique()

Attributes and methods of DataFrames¶

command	description
`df.shape`	number of rows and columns
`df.height`	number of rows
`df.width`	number of columns
`df.schema`	shows data types
`df.head(n)`	shows first n lines
`df.tail(n)`	shows last n lines
`df.glimpse()`	quick summary
`df[col].value_counts()`	counts unique values in a column
`df[col].unique()`	shows unique values in a colunmn
`df.estimated_size()`	returns size in bytes
`df.n_chunks()`	returns number of memory chunks

Challenge¶

Inspect the report stations.csv. Solve the following tasks

display the number of rows and columns
display the last 5 rows
list the column names
how many stations does the ship have?
how many different security levels are there?
which of the stations is under construction?
where did Ming Ming, the intern, go?