one approach is to generate a numerical id variable, naudifirm, for each audifirm, then use
egen x=std(naudifirm), by(company)

if x!=0 the company has different audifirm across year, otherwise, has the same audifirm.

十分感谢，太强了

本来就有变量code区别公司的，但一个公司每年度审计的公司（audfirm）可能不同，用你给的命令还是没有办法判定这个公司t-1年与t年的审计公司不同为1，相同为0啊，亲
请再次执教，谢谢

审计公司是字符型数据，包括200多家公司的名称，code是数值型数据

if the standard deviation of audfirm is 0 for a a company across years, the company has the same audfirm.  Do you want to see if the company has the same audfirm between two consecutive years or across all years?  To see by consecutive years, do

sort company year
gen x=1 if audifirm!=audifirm[_n-1] & company==company[_n-1]
replace x=0 if x==.