Python Projects for Beginners a ten-Week Bootcamp Approach to Python Programming

Subsetting

When you need to filter out records but retain the information within the DataFrame you need to use a concept called subsetting. We’ll use the same condition as earlier, except this time we’ll use it to filter out records rather than create a true-false representation:

# using subsetting to filter out records and keep DataFrame intact df[ df["ages"] > 21 ]

Go ahead and run the cell. The output results in only those records whose ages are equal to or above the age of 21. We took the condition from above and wrapped it within brackets while accessing the df variable. Although it may look weird, the syntax representation is the following:
>>> dataframe_variable [ conditional statement to filter records with ]
You could also write the following for the same exact result:
>>> df[ can_drink ]
Remember that can_drink is a representation of true-false values, which means that the preceding statement will filter out all records that have the value of false.

Column Transformations

Rarely, if ever, will the columns in the original raw DataFrame imported from CSV or a database be the ones you need for your analysis. You will spend lots of time constantly transforming columns or groups of columns using general computational operations to produce new ones that are functions of the old ones. Pandas has full support for this and does it efficiently.

Generating a New Column with Data

To create a new column within a DataFrame, you use the same syntax as if you were adding a new key-value pair into a dictionary. Let’s create a column of fake data that represents how long the people within our DataFrame have been customers with our company:

1| # generating a new column of fake data for each record in the DataFrame to represent customer tenure
2| random.seed(321)
4| tenure = [ random.randint(0, 10) for x in range( len(df) )]
6| df["tenure"] = tenure # same as adding a new key-value pair in a dictionary
7| df.head( )

Go ahead and run the cell. The output will result in a new column created with random numbers for their tenure. We were able to add the column and its values on line 6. In Table 10-2, you’ll find the updated DataFrame, sorted by age.
Table 10-2. Adding a new column to the DataFrame