如何使用「pd.eval」評估 Pandas 中的算術表達式？

Dynamically Evaluating Expressions from Formulas in Pandas

Challenge

Evaluate arithmetic expressions using pd.eval on one or more DataFrame columns, as shown in the following example:

x = 5
df2['D'] = df1['A'] + (df1['B'] * x)

pd.eval Functions

pd.eval, df.eval, and df.query are three closely related functions for evaluating expressions in Pandas. Each has its own subtle variations, but they all follow similar syntax rules and feature support.

Supported Features:

• Arithmetic operations
• Comparison operations
• Boolean operations
• List and tuple literals
• Attribute access
• Subscript expressions
• Simple variable evaluation

Syntax Rules:

Expressions must be passed as strings, with the following guidelines:

• Entire expression is a string
• Variables in the global namespace are referenced by their names
• Specific columns are accessed through attribute accessor
• Parentheses can be used to override operator precedence

Key Differences

pd.eval vs. df.eval

• Column Access: pd.eval requires column names with DataFrame indexing, while df.eval allows direct access to column names.
• Expressions with DataFrames: pd.eval is better for dataframe-wide operations, while df.eval operates on specific DataFrames.

df.eval vs. df.query

• Querying vs. Evaluation: df.query evaluates conditional expressions and returns matching rows. df.eval returns the result of the expression itself.
• Convenience: df.query is generally more concise for querying purposes.

Solution

To solve the original challenge using pd.eval:

x = 5
pd.eval("df1.A + (df1.B * x)")

Reassignment

To assign the result of the expression back to df2, use the target parameter:

pd.eval("D = df1.A + (df1.B * x)", target=df2)

Passing Arguments Inside the Expression

To pass x as an argument within the expression string, use the @ symbol:

pd.eval("df1.A + (df1.B * @x)", local_dict={'x': x})

以上是如何使用「pd.eval」評估 Pandas 中的算術表達式？的詳細內容。更多資訊請關注PHP中文網其他相關文章！