Convert all elements of a matrix into respective inverse values

[Sophie Nguyen]

Hi everyone,

I am having a problem with converting all elements of a matrix to their respective inverse values (NOT the inverse of a matrix). For example, I want element a to become 1/a, element b to become 1/b..... In my example below, the matrix is 8x8. I want to convert all 64 (8*8) elements to their respective inverse values (1/....), so that I can have a new matrix. My real matrix is 63x63 (which is too big for command dataex). I really appreciate your help. Thank you so much!

* Example generated by -dataex-. To install: ssc install dataex
clear
input double p01 int(p02 p04) double(p06 p08) int(p10 p11 p12)
  0 281 269 155 128 287 436 384
284   0 249 252 154 199 427 303
277 228   0 114 240 444 721 570
164 252 114   0 126 330 553 426
130 154 240 126   0 204 424 310
289 199 518 330 204   0 303 101
437 427 722 590 424 301   0 200
386 303 539 426 310 101 202   0
end

[William Lisowski]

It is not clear to me if by "matrix" you mean just the array of data you presented, or a Stata matrix, or a matrix within the Mata environment.

For the first of these options,

. foreach v of varlist * {
  2.     quietly replace `v' = 1/`v'
  3. }

. list, clean noobs

          p01       p02       p04         p06         p08       p10       p11       p12  
            .   .003559   .003717   .00645161    .0078125   .003484   .002294   .002604  
    .00352113         .   .004016   .00396825   .00649351   .005025   .002342     .0033  
    .00361011   .004386         .   .00877193   .00416667   .002252   .001387   .001754  
    .00609756   .003968   .008772           .   .00793651    .00303   .001808   .002347  
    .00769231   .006494   .004167   .00793651           .   .004902   .002358   .003226  
    .00346021   .005025   .001931    .0030303   .00490196         .     .0033   .009901  
    .00228833   .002342   .001385   .00169492   .00235849   .003322         .      .005  
    .00259067     .0033   .001855   .00234742   .00322581   .009901    .00495         .  

. 
​​​​​​​

[Sophie Nguyen]

Thank you so much for your reply, William. I really appreciate.

[Joro Kolev]

See also this FAQ here:
https://www.stata.com/support/faqs/d...s-on-matrices/

[Sophie Nguyen]

Thank you @Joro Kolev!