[Da GXHI]

to Question in #15
Variables 46
Obs 23'465'605
size 2'616'052M. I run the first line again when posting, still not finished

[Nick Cox]

egen is not very fast but on a very mundane laptop this script took about 1 minute to finish. Using Stata 17. Note that most of it is setting up an example dataset, which you don't need to do.

clear 
set rmsg on 
set obs 24000000
egen year = seq(), block(1000000)
set seed 2803 
gen place = runiformint(1, 1000)
gen id = runiformint(1, 1000)

egen tag = tag(place id year)
egen distinct = total(tag), by(place year)

[Da GXHI]

Thanks. What I do is very much the same, just that my id variable is string and my place variable is dummy (1/0). I will give it another try,
Any known reason way tag function may not wor as well with stat 15.1? I am reluctant on updating the stata version as the rest of my code might be affected.

[Nick Cox]

Sorry, but I don't know why this calculation should take hours, as you imply.

[Da GXHI]

use dataset.dta, clear

egen tag_firm = tag(cocode year PC)
egen distinct_firm = total(tag_firm), by(year PC)

*cocode is a string, year is floating, PC is dummy
just n case you see something I don't

[Paris Rira]

I have the same question, but a bit different. I have a variable (Companies: NPC_FIC), some are repeated and some are not. Because more often firms have more than one worker in this case Company's ID appears more than once. I need to compute SUM ( observation, mean, std.dev. min and max) of firms.
* Example generated by -dataex-. For more info, type help dataex
clear
input double NPC_FIC
500000001
500000001
500000001
500000002
500000002
500000002
500000002
500000002
500000033
500000033
500000050
500000050
500000083
500000104
500000119
500000119
500000119
500000119
500000119
500000119
500000119
500000119
500000121
500000121
500000156
500000156
500000157
500000157
500000157
500000157
500000165
500000165
500000165
500000180
500000198
500000198
500000198
500000201
500000204
500000204
500000212
500000215
500000215
500000215
500000215
500000240
500000240
500000276
500000279
500000284
500000284
500000284
500000284
500000305
500000334
500000341
500000341
500000346
500000346
500000346
500000376
500000376
500000395
500000395
500000395
500000395
500000395
500000395
500000465
500000465
500000465
500000470
500000470
500000478
500000525
500000543
500000543
500000565
500000583
500000633
500000633
500000675
500000709
500000748
500000757
500000765
500000765
500000765
500000766
500000766
500000774
500000791
500000791
500000795
500000795
500000795
500000795
500000796
500000800
500000800
end
[/CODE]
------------------ copy up to and including the previous line ------------------


any ideas?

[Sanika Rr]

Hello all,
I am using Stata 16. In my panel data, and I have variables state and district. Each state has a certain number of districts. I want to count how many districts are there per state.

I used the following

egen tag = (district state)

followed by

egen distinct = total(district), by (state)

however it gives me an error
type mismatch

Kindly suggest what mistake I am making here. Thank you.

[Nick Cox]

 
 egen tag = tag(district state)