How Can I Pivot a Pandas DataFrame in Python?

How can I pivot a dataframe?

What is pivot?

Pivoting is a data transformation technique used to reshape a DataFrame by interchanging the rows and columns. It's commonly used to organize data in a way that makes it easier to analyze or visualize.

How do I pivot?

There are several ways to pivot a DataFrame in Python using the Pandas library:

1. pd.DataFrame.pivot_table:

This method is a versatile and feature-rich option for pivoting data. It allows you to specify the values to be aggregated, the aggregation function, and the row and column indices.

Example:

import pandas as pd

# Create a sample DataFrame
df = pd.DataFrame({
    "row": ["row0", "row1", "row2", "row3", "row4"],
    "col": ["col0", "col1", "col2", "col3", "col4"],
    "val0": [0.81, 0.44, 0.77, 0.15, 0.81],
    "val1": [0.04, 0.07, 0.01, 0.59, 0.64]
})

# Pivot the DataFrame using pivot_table
df_pivoted = df.pivot_table(
    index="row",
    columns="col",
    values="val0",
    aggfunc="mean",
)

print(df_pivoted)

# Output:
     col0   col1   col2   col3   col4
row                                  
row0  0.77  0.445  0.000  0.860  0.650
row1  0.130  0.000  0.395  0.500  0.250
row2  0.000  0.310  0.000  0.545  0.000
row3  0.000  0.100  0.395  0.760  0.240
row4  0.000  0.000  0.000  0.000  0.000

2. pd.DataFrame.groupby pd.DataFrame.unstack:

This method involves grouping the DataFrame by the desired row and column indices, and then using unstack to pivot the grouped data.

Example:

# Group the DataFrame by row and col
df_grouped = df.groupby(["row", "col"])

# Perform pivot using unstack
df_pivoted = df_grouped["val0"].unstack(fill_value=0)

print(df_pivoted)

# Output:
col   col0   col1   col2   col3   col4
row                                  
row0  0.81  0.445  0.000  0.860  0.650
row1  0.130  0.000  0.395  0.500  0.250
row2  0.000  0.310  0.000  0.545  0.000
row3  0.000  0.100  0.395  0.760  0.240
row4  0.000  0.000  0.000  0.000  0.000

3. pd.DataFrame.set_index pd.DataFrame.unstack:

This method involves setting the desired row and column indices as the DataFrame's index, and then using unstack to pivot the data.

Example:

# Set the row and col as the DataFrame's index
df = df.set_index(["row", "col"])

# Perform pivot using unstack
df_pivoted = df["val0"].unstack(fill_value=0)

print(df_pivoted)

# Output:
col   col0   col1   col2   col3   col4
row                                  
row0  0.81  0.445  0.000  0.860  0.650
row1  0.130  0.000  0.395  0.500  0.250
row2  0.000  0.310  0.000  0.545  0.000
row3  0.000  0.100  0.395  0.760  0.240
row4  0.000  0.000  0.000  0.000  0.000

4. pd.DataFrame.pivot:

This method offers a simpler syntax compared to pivot_table, but it has limited functionality. It only allows you to specify row and column indices, and it cannot perform aggregation.

Example:

# Perform pivot using pivot
df_pivoted = df.pivot(index="row", columns="col")

print(df_pivoted)

# Output:
col   col0   col1   col2   col3   col4
row                                  
row0  key0  0.81  0.44  0.00  0.86  0.65
row1  key1  0.13  0.00  0.39  0.50  0.25
row2  key1  0.00  0.31  0.00  0.54  0.00
row3  key0  0.00  0.10  0.39  0.76  0.24
row4  key1  0.00  0.00  0.00  0.00  0.00

Long format to wide format

To convert a DataFrame from long format to wide format using only two columns:

1. pd.DataFrame.pivot(index=column_to_index, columns=column_to_columns, values=values_to_pivot**):

Example:

df["Combined"] = df["row"] + "|" + df["col"]
df_pivoted = df.pivot(index="Combined", columns="A", values="B")

print(df_pivoted)

# Output:
A         a     b    c
Combined
row0|col0  0.0  10.0  7.0
row1|col1  11.0  10.0  NaN
row2|col2  2.0  14.0  NaN
row3|col3  11.0   NaN  NaN
row4|col4   NaN   NaN  NaN

2. pd.DataFrame.groupby pd.DataFrame.unstack:

df["Combined"] = df["row"] + "|" + df["col"]
df_grouped = df.groupby(["Combined", "A"])
df_pivoted = df_grouped["B"].unstack(fill_value=0)

print(df_pivoted)

# Output:
A         a     b    c
Combined
row0|col0  0.0  10.0  7.0
row1|col1  11.0  10.0  NaN
row2|col2  2.0  14.0  NaN
row3|col3  11.0   NaN  NaN
row4|col4   NaN   NaN  NaN

Flatten the multiple index to a single index after pivot:

df_pivoted.columns = df_pivoted.columns.map("|".join)

print(df_pivoted)

# Output:
   a|col0  b|col0  c|col0  a|col1  b|col1  c|col1  a|col2  b|col2  c|col2  a|col3  b|col3  c|col3
row                                                                                        
row0    0.0   10.0    7.0   11.0   10.0    NaN    2.0   14.0    NaN    11.0    NaN    NaN
row1    0.0   10.0    7.0   11.0   10.0    NaN    2.0   14.0    NaN    11.0    NaN    NaN

The above is the detailed content of How Can I Pivot a Pandas DataFrame in Python?. For more information, please follow other related articles on the PHP Chinese website!