Extracting first five and last three parts from string variable.

[Ed Suh]

Dear Stata users,

I hope you are staying safe.
I have data set with code and name like this.

* Example generated by -dataex-. To install: ssc install dataex
clear
input str12 code str36 name
"160610117117" "A"                  
"160610118118" "B"           
"160615001003" "C"                 
"160615001015" "D"                       
"160615001026" "E"                 
"160615001031" "F"                        
"160615001094" "G"                    
"160615001103" "H"                       
"160615001114" "I"                  
end

Hi, I'd like to change these codes as follows:
160610117117 -> 16061117

So, I want to reformulate the codes with first five and last three parts.

I tried some codes with substr() but could not figure out.. By the way, the codes are string variable! If you think changing to numeric is better let me konw!

[Nick Cox]

gen wanted1 = substr(code, 1, 5) + substr(code, -3, 3)
gen wanted2 = subinstr(code, substr(code, 6, 4), "", 1)

l wanted?

We'd need to see your code to tell you what you did wrong.

[Ed Suh]

Hi Nick,

thank you for the reply.
It is what I was supposed to do.. I did not know
substr(code, -3, 3) is also available like backwards counting. Thank you!

[Joro Kolev]

Yes Ed, backward counting is available in many contexts, including with the -in-qualifier (see toward the end of this example). But this is needed only if the string in the middle has varying length, which does not seem to be the case in your data. For your data forward counting would have done just fine. Nick has already illustrated concatenation with the + operator (which in this context results in a one-liner) so I will do it in the more round about way, which in turn is more handy if you want to operate on general varlists:

. gen first = substr(code,1,5)

. gen last = substr(code,10,13)

. egen shortcode = concat(first last)

. list in -4/l

     +-----------------------------------------------+
     |         code   name   first   last   shortc~e |
     |-----------------------------------------------|
  6. | 160615001031      F   16061    031   16061031 |
  7. | 160615001094      G   16061    094   16061094 |
  8. | 160615001103      H   16061    103   16061103 |
  9. | 160615001114      I   16061    114   16061114 |
     +-----------------------------------------------+

[Nick Cox]

Examples or sources of extra information:

help functions 
help string functions
help substr()

https://www.stata-journal.com/articl...article=dm0058