split and rejoin words

[Jarvin99]

data have;
input ID   rolename$;
cards; 
1      vp/ceo
1      division vp/ceo
1      global ceo/ cfo/ coo
2      vice ceo/cfo
2      vice ceo/cfo
2      division ceo/coo/cfo 
;

I want only strings starting with division and global to change to:

data have;
input ID   rolename$;
cards; 
1      vp/ceo
1      division vp/division ceo
1      global ceo/global cfo/global coo
2      vice ceo/division cfo
2      vice ceo/cfo
2      division ceo/division coo/division cfo 
;

Anyway to do? Eventually I want to detect that as long as the rolename contains "division ceo", the indicator will indicate 0.

Thank you.

[Jarvin99]

Just an add on, if possible I want this split-and-rejoin thing to be a macro, as I actually have more starting words (not just "division" and "global", but also words like "regional" and "international") which I want to conduct this.

[ballardw]

---

@Jarvin99 wrote:

Just an add on, if possible I want this split-and-rejoin thing to be a macro, as I actually have more starting words (not just "division" and "global", but also words like "regional" and "international") which I want to conduct this.

---

I have a strong suspicion that adding in macro language will just make this more complicated. The bit you need to provide are the rules about when a specific value is to be used.

[ballardw]

First thing is to ensure that your posted data step creates the values you expect. Yours does not because the $ creates variables of length 8 and use of list input means that the first space encounters ends the value of "rolename".

Second, you will need to either specify a length for that variable that will allow inserting extra characters.

Here is an example data set that makes rolename long enough to insert some characters but you may have to provide much more exhaustive examples to know just how long it should be.

data have;
length id 8 rolename $ 100;
input ID   rolename 8-28;
cards; 
1      vp/ceo
1      division vp/ceo
1      global ceo/ cfo/ coo
2      vice ceo/cfo
2      vice ceo/cfo
2      division ceo/coo/cfo 
;

Now to heart of the problem.

What in the data tells us that the first occurrence of

vice ceo/cfo

is supposed to become

vice ceo/division cfo

but the second occurrence does not change???

Duplicating a first word separated by a space, which appeared to be a rule at first glance can be done relatively easily:

data want;
   set have;
   if not missing(scan(rolename,1,' ')) then 
        rolename= tranwrd(rolename,'/',cat('/',scan(rolename,1,' '),' '));
run;

but there appear to be a multitude of other rules involved as you insert a different word for some, nothing for others.

[Jarvin99]

Original:

data have;
input ID   rolename$50;
cards; 
1      vp/ceo
1      division vp/ceo
1      global ceo/cfo/coo
2      vice ceo/division cfo
2      vice ceo/cfo
2      division ceo/coo/cfo 
;

Hope to achieve:

data have;
input ID   rolename$50;
cards; 
1      vp/ceo
1      division vp/division ceo
1      global ceo/global cfo/global coo
2      vice ceo/division cfo
2      vice ceo/cfo
2      division ceo/division coo/division cfo 
;

In this case, no macro is fine, but I need to keep the starting word flexible as I have other starting words (e.g., international ceo/cfo --> international ceo/international cfo) to apply.

 Sorry for the data input issue as I am not very familiar with this. Thank you so much for your help. 

[Jarvin99]

Not all starting words I want to split and rejoin. In this case, only "division" and "global" I want to do this, but not for "vice". Thus, the issue is a bit troublesome.

[Ksharp · Accepted solution]

data have;
input ID   rolename $50.;
cards; 
1      vp/ceo
1      division vp/ceo
1      global ceo/cfo/coo
2      vice ceo/division cfo
2      vice ceo/cfo
2      division ceo/coo/cfo 
;
data want;
 set have;
 length want $ 200;
 want=rolename;
 if rolename =: 'division' then want=prxchange('s/\//\/division /o',-1,rolename);
 if rolename =: 'global'   then want=prxchange('s/\//\/global /o',-1,rolename);
run;