The command tabstat can generate coefficient of variation estimates by a single group.
How could I calculate the coefficient of variation for two groups?
As if the following was allowed in Stata:
tabstat price, statistics(n mean cv) by(group_var1 group_var2)
You can calculate and display it manually:
bysort group_var1 group_var2 : egen sd = sd(price)
bysort group_var1 group_var2 : egen mean = mean(price)
generate cv = sd / mean
tabdisp group_var1 group_var2, c(cv)
You can optionally scale by 100, if you like.