How to create unique identifiers (IDs) for each observation

[Victor Smith]

Hi there, I am analyzing some data at the dyadic level, and would like to create some identifiers columns to represent the actors, the partners, and their dyads.

The raw data table goes like this:

teamID	actorName	partnerName
1	Adam	Raddi
1	Adam	Samatha
1	Adam	JoJo
1	Raddi	Adam
1	Raddi	Samatha
1	Raddi	JoJo
1	Samatha	Adam
1	Samatha	Raddi
1	Samatha	JoJo
1	JoJo	Adam
1	JoJo	Raddi
1	JoJo	Samatha
2	Nix	Kim
2	Nix	Susan
2	Kim	Nix
2	Kim	Susan
2	Susan	Nix
2	Susan	Kim

I would like to have a table with the ID columns as the one below. there are some requirements for the ID columns.

1. Each actor or partner should have a unique ID.

2. The actor_ID should correspond with the partner_ID. For example, when Adam is the "actor", his actor_ID is "1", thus, when Adam becomes the "partner", his partner_ID should still be "1". This example is highlighted in red below. The example in red.

3. Each dyad should have a unique ID regardless of the order of actor and partner. That is, for the dyad of Adam and Raddi, the dyad_ID is "1" no matter (Adam is the actor and Raddi is the partner) or (Raddi is the actor and Adam is the partner). This example is highlighted in green.

Does anyone know how to generate these ID columns?

teamID	actorName	partnerName	actor_ID	patner_ID	dyad_ID
1	Adam	Raddi	1	2	1
1	Adam	Samatha	1	3	2
1	Adam	JoJo	1	4	3
1	Raddi	Adam	2	1	1
1	Raddi	Samatha	2	3	4
1	Raddi	JoJo	2	4	5
1	Samatha	Adam	3	1	2
1	Samatha	Raddi	3	2	4
1	Samatha	JoJo	3	4	6
1	JoJo	Adam	4	1	3
1	JoJo	Raddi	4	2	5
1	JoJo	Samatha	4	3	6
2	Nix	Kim	5	6	7
2	Nix	Susan	5	7	8
2	Kim	Nix	6	5	7
2	Kim	Susan	6	7	9
2	Susan	Nix	7	5	8
2	Susan	Kim	7	6	9

[Nick Cox]

Good news: all previously discussed here.

To get consistent numeric identifiers for your name variables, use multencode from SSC, as first announced here

https://www.stata.com/statalist/arch.../msg00729.html

and discussed several times since the forum started, most recently in

https://www.statalist.org/forums/for...es-with-encode

The general territory of identifiers for dyads was discussed in http://www.stata-journal.com/sjpdf.h...iclenum=dm0043

that is

SJ-8-4 dm0043 . Tip 71: The problem of split identity, or how to group dyads
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . N. J. Cox
Q4/08 SJ 8(4):588--591 (no commands)
tip on how to handle dyadic identifiers


and dm0043 is thus revealed as an otherwise unpredictable search term to find related discussions here.

That said, how does this work on your example?

* Example generated by -dataex-. To install: ssc install dataex
clear
input byte teamid str7(actorname partnername)
1 "Adam"    "Raddi"  
1 "Adam"    "Samatha"
1 "Adam"    "JoJo"  
1 "Raddi"   "Adam"  
1 "Raddi"   "Samatha"
1 "Raddi"   "JoJo"  
1 "Samatha" "Adam"  
1 "Samatha" "Raddi"  
1 "Samatha" "JoJo"  
1 "JoJo"    "Adam"  
1 "JoJo"    "Raddi"  
1 "JoJo"    "Samatha"
2 "Nix"     "Kim"    
2 "Nix"     "Susan"  
2 "Kim"     "Nix"    
2 "Kim"     "Susan"  
2 "Susan"   "Nix"    
2 "Susan"   "Kim"    
end

* type "ssc install multencode" to install
multencode actorname partnername, gen(actor_id partner_id)

gen dyad_id = cond(actor_id < partner_id, string(actor_id) + " " + string(partner_id), string(partner_id) + " " + string(actor_id))
egen dyad = group(dyad_id), label

list, sepby(teamid)

     +---------------------------------------------------------------------+
     | teamid   actorn~e   partne~e   actor_id   partne~d   dyad_id   dyad |
     |---------------------------------------------------------------------|
  1. |      1       Adam      Raddi       Adam      Raddi       1 5    1 5 |
  2. |      1       Adam    Samatha       Adam    Samatha       1 6    1 6 |
  3. |      1       Adam       JoJo       Adam       JoJo       1 2    1 2 |
  4. |      1      Raddi       Adam      Raddi       Adam       1 5    1 5 |
  5. |      1      Raddi    Samatha      Raddi    Samatha       5 6    5 6 |
  6. |      1      Raddi       JoJo      Raddi       JoJo       2 5    2 5 |
  7. |      1    Samatha       Adam    Samatha       Adam       1 6    1 6 |
  8. |      1    Samatha      Raddi    Samatha      Raddi       5 6    5 6 |
  9. |      1    Samatha       JoJo    Samatha       JoJo       2 6    2 6 |
 10. |      1       JoJo       Adam       JoJo       Adam       1 2    1 2 |
 11. |      1       JoJo      Raddi       JoJo      Raddi       2 5    2 5 |
 12. |      1       JoJo    Samatha       JoJo    Samatha       2 6    2 6 |
     |---------------------------------------------------------------------|
 13. |      2        Nix        Kim        Nix        Kim       3 4    3 4 |
 14. |      2        Nix      Susan        Nix      Susan       4 7    4 7 |
 15. |      2        Kim        Nix        Kim        Nix       3 4    3 4 |
 16. |      2        Kim      Susan        Kim      Susan       3 7    3 7 |
 17. |      2      Susan        Nix      Susan        Nix       4 7    4 7 |
 18. |      2      Susan        Kim      Susan        Kim       3 7    3 7 |
     +---------------------------------------------------------------------+

 list, sepby(teamid) nolabel

     +---------------------------------------------------------------------+
     | teamid   actorn~e   partne~e   actor_id   partne~d   dyad_id   dyad |
     |---------------------------------------------------------------------|
  1. |      1       Adam      Raddi          1          5       1 5      2 |
  2. |      1       Adam    Samatha          1          6       1 6      3 |
  3. |      1       Adam       JoJo          1          2       1 2      1 |
  4. |      1      Raddi       Adam          5          1       1 5      2 |
  5. |      1      Raddi    Samatha          5          6       5 6      9 |
  6. |      1      Raddi       JoJo          5          2       2 5      4 |
  7. |      1    Samatha       Adam          6          1       1 6      3 |
  8. |      1    Samatha      Raddi          6          5       5 6      9 |
  9. |      1    Samatha       JoJo          6          2       2 6      5 |
 10. |      1       JoJo       Adam          2          1       1 2      1 |
 11. |      1       JoJo      Raddi          2          5       2 5      4 |
 12. |      1       JoJo    Samatha          2          6       2 6      5 |
     |---------------------------------------------------------------------|
 13. |      2        Nix        Kim          4          3       3 4      6 |
 14. |      2        Nix      Susan          4          7       4 7      8 |
 15. |      2        Kim        Nix          3          4       3 4      6 |
 16. |      2        Kim      Susan          3          7       3 7      7 |
 17. |      2      Susan        Nix          7          4       4 7      8 |
 18. |      2      Susan        Kim          7          3       3 7      7 |
     +---------------------------------------------------------------------+

The numbers are not all the same, but the spirit of identifiers is what I think you're asking for.

[Victor Smith]

Thanks Nick!