Cleaning string variable

[marco lecci]

Dear all,

I have a variable called place_birth in my dataset. Some of the locations weren't recorded properly.

place_birth

Feucherolles (Saint-James = Le château royal de Sainte-Gemme)
ST B(?), canton de Chaillot
(?Chanvrand)Canton de La Guiche
Seine-Inférieure (Seine-Maritime)
Épinay-sur-Seine ,
Autine (?) Outines
Darrois ? Darvois

I would like to do two things.
First, separate what is inside parenthesis () and comma , and = from the text. With what I separate I can create an new variable called place_new
Second, clean both variable from weird signs like ?, =, . at the end, /, etc...

For example
Épinay-sur-Seine ,

should look like
Épinay-sur-Seine

replace ? and (?) with a comma
Autine (?) Outines
it becomes
Autine , Outines

For this one:
Feucherolles (Saint-James = Le château royal de Sainte-Gemme)

Eliminate "Saint-James =" and just leave:
Feucherolles (Le château royal de Sainte-Gemme)

Then I can separate the strings by comma and parenthesis so that for example:

place_birth
(?Chanvrand)Canton de La Guiche

becomes:
place_new
Chanvrand

Or:
place_birth
Seine-Inférieure (Seine-Maritime)

Becomes in the new var:
place_new
Seine-Maritime

[Hemanshu Kumar]

marco lecci you essentially need a bunch of manipulations using functions like subinstr() and commands like split. You can look up the help for these, try them out, and come back to the forum if you're stuck.

[marco lecci]

Thanks Hemanshu Kumar
I did try
gen dep_new = ustrregexs(0) if ustrregexm(départementdenaissance,"((?<=\().*?(?=\ )))|(^(^\(\))*$)")
This allows me to split the variable into (). However, the cleaning for ?, commas etc... I don't get the results I want.