DataFrames

Question 1

Array

Answer 1

• one-dimensional
• unordered collection
• contains only one data type

Each item has an index and a value.

Question 2

Three main components of tables

Answer 2

rows, columns, index

Question 3

Rows of a table

Answer 3

“entry” or “observation”

Question 4

Columns of a table

Answer 4

Each column of a table represents some attribute that entries (rows) have.

Question 5

Index of a table

Answer 5

• The first column
• Meaningful or arbitrary
• Unique values
• Identify rows

Question 6

Series

Answer 6

• The most basic pandas object
• Has two sections: the index and the values
• Under the hood, columns of a Series are actually NumPy arrays

Question 7

DataFrames

Answer 7

• Pandas table object
• Contains an Index, Rows, and Columns
• Each column is a Series

Question 8

How do you read a DataFrame?

Answer 8

pd.read_csv(filepath)

Question 9

df.loc[]

Answer 9

• Accesses rows/columns by label
• Loc slicing is right-inclusive
• Syntax: df.loc[A:B, C:D]

Question 10

filtering using df.loc[]

Answer 10

Ex: movies.loc[movies[“Year”] < 1950]

• You can also filter by more than one condition using
  condition1 = movies[“Year”] >= 2000
  condition2 = movies[“Studio”] == “Fox”

filtered_or = movies.loc[condition1 | condition2]

filtered_and = movies.loc[condition1 & condition2]

Question 11

How do you assign columns to a DataFrame?

Answer 11

Using indexing/loc:
- df[“column”] data
Using df.assign():
- new_df = df.assign(label=data)

Question 12

How do you sort DataFrames?

Answer 12

df.sort_values()

Ex: movies.sort_values(“Studio”, ascending=True)
Ex: movies.sort_values(“Year”, ascending=False)

Question 13

df.groupby()

Answer 13

Creates new df grouped by certain column(s)
Ex: df.groupby([col1, col2, …])

Question 14

Ways of grouping by 2 columns

Answer 14

1. using df.groupby()
   Ex: movies.groupby([“Year”, “Studio”])[“Title”].count().to_frame()
2. using df.pivot_table()
   Ex: pt = movies.pivot_table(values=”Title”, index=”Year”, columns=”Studio”, aggfunc=”count”)

Question 15

Merging DataFrames

Answer 15

pd.merge()
Inner Join:
- This will only include rows with a match in both DataFrames.
Ex: pd.merge(adf, bdf, how=”inner”, on=”x1”)
Outer Join:
- This will retain all rows in both DataFrames.
Ex: pd.merge(adf, bdf, how=”outer”, on=”x1”)
Left Join:
Use all rows form the First DataFrame
Right Join:
Use all rows from the second DataFrame

Question 16

df.apply()

Answer 16

Applies a function to a DataFrame
Applying to a column:
- Will return a Series with the function applied to each row in the column
df[column].apply(function)
It is possible to apply a function to a row with axis=1
df.apply(function, axis=1)