变量
在分析的过程中，有些变量并没有在数据中提供，需要我们用原始数据或者回归的结果构造。常用的命令是 “gen” 和 “egen” 。

格式
gen educsqr=educ^2

egen命令相对复杂一些，它能生成一些“gen”命令无法生成的变量。例如可以生成wagesum为每个人的工资和，以及生成wagemedian为工资的中位数(median)，wagemax为工资的最大值。

格式:
egen wagesum=sum(wage)
egen wagemedian=median(wage)
egen wagemax=max(wage)

更复杂的如想产生一个变量“wagemax”为相同教育水平里的最高工资。

格式: egen wagemaxeduc=max (wage),by (educ)

egen 和 gen 都用于生成新变量，但egen 的特点是它更强大的函数功能。 gen 可以支持一些函数， egen 支持额外的函数。如果用 gen 搞不定，就得用egen想办法了。不知道有没有说明白？

Hello, the egen is an excellent stata command which is useful in particular for large database for which variables contain repeated sequence. It's used in general with other commands (by (id), ...).
You can see this document from stata journal 2006 for explanation and exemples.

i once had the same question. here is what i did.

first, i get all the egen commands--including those from "egenmore" (additional egen commands).
then, try those commands with some data points.

good!