removing numbers from string

[Raphael George]

Hi, I want to create three separate variables with the numbers between the letters. For instance from c101r10r145 I want to create three variables namely 'district' with value=145 , state=10 and country=101, how do I separate them?

----------------------- copy starting from the next line -----------------------


	
	
* Example generated by -dataex-. To install: ssc install dataex
clear
input str12 ID_adm2
"c100r0r0"   
"c101r10r144"
"c101r10r145"
"c101r10r146"
"c101r10r147"
"c101r10r148"
"c101r10r149"
"c101r10r150"
"c101r10r151"
"c101r10r152"
"c101r10r153"
"c101r10r154"
"c101r10r155"
"c101r10r156"
"c101r10r157"
"c101r10r158"
"c101r10r159"
"c101r10r160"
"c101r11r161"
"c101r11r162"
"c101r11r163"
"c101r11r164"
"c101r12r165"
"c101r12r166"
"c101r12r167"
"c101r12r168"
"c101r12r169"
"c101r12r170"
"c101r12r171"
"c101r12r172"
"c101r12r173"
"c101r12r174"
"c101r12r175"
"c101r12r176"
"c101r12r177"
"c101r12r178"
"c101r12r179"
"c101r12r180"
"c101r12r181"
"c101r12r182"
"c101r12r183"
"c101r13r184"
"c101r13r185"
"c101r13r186"
"c101r13r187"
"c101r13r188"
"c101r13r189"
"c101r13r190"
"c101r13r191"
"c101r13r192"
"c101r13r193"
"c101r13r194"
"c101r13r195"
"c101r13r196"
"c101r13r197"
"c101r13r198"
"c101r13r199"
"c101r13r200"
"c101r13r201"
"c101r13r202"
"c101r13r203"
"c101r13r204"
"c101r13r205"
"c101r13r206"
"c101r13r207"
"c101r13r208"
"c101r13r209"
"c101r13r210"
"c101r13r211"
"c101r14r212"
"c101r14r213"
"c101r14r214"
"c101r14r215"
"c101r14r216"
"c101r14r217"
"c101r14r218"
"c101r14r219"
"c101r14r220"
"c101r14r221"
"c101r14r222"
"c101r14r223"
"c101r14r224"
"c101r14r225"
"c101r14r226"
"c101r14r227"
"c101r15r228"
"c101r15r229"
"c101r15r230"
"c101r15r231"
"c101r15r232"
"c101r15r233"
"c101r15r234"
"c101r15r235"
"c101r15r236"
"c101r15r237"
"c101r15r238"
"c101r15r239"
"c101r15r240"
"c101r15r241"
"c101r15r242"
end
--------

[Andrew Musau]

With just a basic understanding of regular expressions, you can accomplish a great deal.

replace ID_adm2= ustrregexra(ID_adm2, "[\d]", "")

[Andrew Musau]

I'm responding to your title in #2.

For instance from c101r10r145 I want to create three variables namely 'district' with value=145 , state=10 and country=101, how do I separate them?

forval i=1/3{
    gen wanted`i'=ustrregexra(ID_adm2, "c(.*)r(.*)r(.*)", "$`i'")
}

[Daniel Feenberg]

Or possibly

gen x=substr(ID_adm2,2,3)
gen y=substr(ID_adm2,6,2)
gen z=substr(ID_adm2,9,3)

if the all the values are as regular as they appear in the sample. See https://www.stata.com/manuals/m-5substr.pdf for -substr-.

[Nick Cox]

Also

. split ID_adm2, parse(r c) destring 
variables born as string: 
ID_adm21  ID_adm22  ID_adm23  ID_adm24
ID_adm21: all characters numeric; replaced as byte
(100 missing values generated)
ID_adm22: all characters numeric; replaced as int
ID_adm23: all characters numeric; replaced as byte
ID_adm24: all characters numeric; replaced as int

. 
. su ID_adm2? 

    Variable |        Obs        Mean    Std. dev.       Min        Max
-------------+---------------------------------------------------------
    ID_adm21 |          0
    ID_adm22 |        100      100.99          .1        100        101
    ID_adm23 |        100       12.55     2.04186          0         15
    ID_adm24 |        100      191.07    34.48415          0        242