Stata中如何生成命名为d+原变量名的滞后差分项变量？

To achieve your goal—creating new difference variables prefixed with d instead of overwriting the original numeric variables—you just need to adjust the variable name in the gen command within your loop. Here's the modified code:

// Get all numeric variables in the dataset
ds, has(type numeric)

// Loop through each numeric variable and create a new difference variable
foreach var in `r(varlist)' {
    gen d`var' = `var' - `var'[_n-1]
}

Key Details: #

• The ds, has(type numeric) command still collects all numeric variables into the macro r(varlist), just like your original code.
• In the loop, dvar'constructs the new variable name by prependingdto the original variable name (e.g., if the original variable isincome, the new one becomes dincome`).
• The first observation of each new dvar' variable will be missing (.), since there’s no prior observation (_n-1` doesn’t exist for the first row)—this is normal behavior for this kind of difference calculation.

Important Note: #

• Ensure your dataset is sorted correctly before running this code. The _n-1 reference depends on the current row order. For time-series or panel data, sort by the relevant identifier and time variable first (e.g., sort id year) to guarantee the difference is computed across consecutive observations as intended.

内容的提问来源于stack exchange，提问作者Tom