How to retrieve a specific value of a variable and calculate it with other values of variables?

[Geralt Ji]

Hi Statalists,

I'm currently want to let stata calculate P1 for each id and here is a formula: P1 = A11 * D1 + ... + A18 * D4iii
(A11 is the first value of A1 which equals 0.345865; A12 is the second value of A1 which equals 0.326931 ; A13 is the third one and so on)

For example, if I want to get the P1 for first person, here is the calculation:
0.345865 (A11) * 1.566901 (D1 for ID == 1) + 0.326931 (A12) * 0.233873 (D2i for ID == 1) ..... + 0.321616 (A18) * 1.384699 (D4iii)

I find it laborious and I wonder if there is any more convenient way to calculate P1 for other IDs automatically?

If I also want to calculate P2 for first person after calculating P1, P2 = A21 * D1 + .... + A28 * D4iii, do I need to use some syntax like for loop?

id	D1	D2i	D2ii	D3i	D3ii	D4i	D4ii	D4iii	A1	A2
1	1.566901	0.233873	0.286958	0.556992	1.113236	1.567941	1.490475	1.384699	0.345865	-0.07561
2	-0.39339	-0.64672	0.07849	-0.88745	-0.2885	-0.57468	-0.70157	-0.59344	0.326931	0.470004
3	0.390256	1.464312	1.263042	0.840091	0.97695	-0.14107	-0.23965	-0.59344	0.320997	0.476968
4	-0.69505	-1.16996	-1.51789	-1.27133	-1.12446	-1.06964	-1.0088	-0.92313	0.367527	0.146237
5	-0.86872	0.118493	-0.1106	0.761692	-0.67723	0.217444	0.459544	0.725319	0.376543	0.223233
6									0.387949	-0.32069
7									0.373703	-0.36884
8									0.321616	-0.48557

[William Lisowski]

I think this code does what you want.

* Example generated by -dataex-. For more info, type help dataex
clear
input byte d float(d1 d2i d2ii d3i d3ii d4i d4ii d4iii a1 a2)
1 1.566901  .233873  .286958  .556992 1.113236 1.567941 1.490475 1.384699 .345865 -.07561
2  -.39339  -.64672   .07849  -.88745   -.2885  -.57468  -.70157  -.59344 .326931 .470004
3  .390256 1.464312 1.263042  .840091   .97695  -.14107  -.23965  -.59344 .320997 .476968
4  -.69505 -1.16996 -1.51789 -1.27133 -1.12446 -1.06964  -1.0088  -.92313 .367527 .146237
5  -.86872  .118493   -.1106  .761692  -.67723  .217444  .459544  .725319 .376543 .223233
6        .        .        .        .        .        .        .        . .387949 -.32069
7        .        .        .        .        .        .        .        . .373703 -.36884
8        .        .        .        .        .        .        .        . .321616 -.48557
end

local vars d1 d2i d2ii d3i d3ii d4i d4ii d4iii
generate p1 = 0
generate p2 = 0
forvalues v=1/8 {
    local var : word `v' of `vars'
    quietly replace p1 = p1 + a1[`v']*`var'
    quietly replace p2 = p2 + a2[`v']*`var'
}
list d p1 p2, clean

. list d p1 p2, clean

       d          p1          p2  
  1.   1    2.945018   -1.266657  
  2.   2   -1.433077    .3002576  
  3.   3    1.360612    2.023885  
  4.   4   -3.089634   -.4948973  
  5.   5    .2170774   -.5625926  
  6.   6           .           .  
  7.   7           .           .  
  8.   8           .           .

Please note the use of the dataex command to provide example data in an immediately usable format.

[Geralt Ji]

Thanks William! It works perfectly.
I'm not familiar with this command. Would you mind telling me more details about what's the function of "word" and what does a1[`v']*`var' mean? Why a1 is attached to [`v']?

[William Lisowski]

See the output of

help local

in the section "Macro functions for parsing" for an explanation of the word function, and see the output of

help subscript

for an explanation of the use of subscripts to select the value of a variable from a particular observation.