I'm trying to understand how to interpret the i(varlist) in reshape when it is a grouped variable created from egen's group() option. I recently created a number of new variables using collapse, e.g. mean, sd, count (by(intra agegrp wave) was advised that both agegrp and wave combined (egen i = group(agegrp wave)) would be a unique identifier - but I am not yet clear how this (as a grouped variable) uniquely identifies the observations.

Code:
list x_totasset1 sd_totasset1 agegrp wave i in 1/10
   +------------------------------------------+
     | x_tota~1   sd_to~t1   agegrp   wave    i |
     |------------------------------------------|
  1. |    33198   23337.35       20      2    1 |
  2. |    37311   65327.32       20      6    2 |
  3. | 120297.4   199242.6       20     10    3 |
  4. | 174395.2   279109.5       20     14    4 |
  5. | 439030.5   726204.7       20     18    5 |
     |------------------------------------------|
  6. | 124624.2   180614.7       25      2    6 |
  7. | 149154.2   159179.2       25      6    7 |
  8. | 226977.2   271007.8       25     10    8 |
  9. | 329907.3   928040.2       25     14    9 |
 10. | 211385.7     440034       25     18   10 |
     +------------------------------------------+
I then apply this to:
Code:
reshape wide x_* sd_* n_* ub_* lb_*, i(i) j(intra)
My best guess is that each of the variables for which I created the mean, sd and count, which are ordered by agegrp and wave will be dropped into each category of the variable intra - e.g. where intra=1, by order of agegroup, then wave, I will have the mean, sd and count of each variable in my varlist. Is that correct?

(related) I did read the example in this Stata FAQ, where Nick Cox shows that when moving from wide to long that id (nor id and pos combined), uniquely identified the observations, so created a new ID and created a new identifier once in long format by using egen to group(id pos) but this didn't help me understand the above example.

Stata v.15.1. Using panel data.