Keeping both " and ' in compress

[Emma_at_SAS]

I have a character variable that contains alphabets, numbers, some symbols, and both ' and " and also some strange characters (characters that are not on any keyboard). I want to create a new variable with only the characters I want. Because I am using " before and after my list, I cannot write " to keep it in my observations for the variable var. How may I include "?

Thanks!

data want;
set have;
new_var=compress(var,"ABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890$%.-':|&[]{}*!@#()!/+,");
run;

[Tom]

To include the character used to quote a string literal just double it up in the string content. Example:

word = 'Don''t';

So if you are using " on the outside to include it as one of the characters just add "" to the list.

compress(var,"ABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890$%.-""':|&[]{}*!@#()!/+,")

[ErikLund_Jensen]

Hi @Emma_at_SAS 

It might be easier to use the "k" modifier. It reverses the compress, so only the specifies characters are kept indtead of dropped.

Drop all printable and non-printable chars except "ABC":   

compress(str,'ABC','k');

Drop all printable and non-printable chars except digits, english chars and underscore. There are no characters specified, but the mentioned chars are added with the 'n' modifier. More examples in the online documentation: 

compress(str,,'nk');

[Emma_at_SAS]

Thank you, @Tom and @ErikLund_Jensen. A combination of your suggestions solved my problem

I used the double "" and,"k" modifier and that gives me exactly what I want.

Thanks!

[Emma_at_SAS]

May someone please help me to mark this conversation as solved? I am not sure because it was solved based on two suggestions.

Thanks

[novinosrin]

How about just-

compress(var,,'p');

[Emma_at_SAS]

Thank you, novinosrin. What does this combination do? Thanks!

[Amir]

Hi @Emma_at_SAS,

From the documentation of the scan() function, when the modifier argument is set to "p", as in @novinosrin's post, punctuation is added to the list of characters that are to be removed from the text:

p or P

adds punctuation marks to the list of characters.

Kind regards,

Amir.

[Ksharp]

new_var=compress(var, , 'kg');

[Emma_at_SAS]

data test;
input var$ ;
datalines;
oÔÇÖ\[}s 
dsd_'"sf 
;
run;

Obs	var
1	oÔÇÖ\[}s
2	dsd_'"sf

data want;
set test;
new_var=compress(var,"ABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890$%.-'"":|&[]{}*!@#()!/+, ",'kg');
run;

proc print data=want; run;

The SAS System

Obs	var	new_var
1	oÔÇÖ\[}s	oÔÇÖ\[}s
2	dsd_'"sf	dsd_'"sf

 Thank you @Ksharp for your suggestion. I already have a solution but I thought this might be easier to just remove the "graphic characters". Beacuse I want to keep everything except the translation of emojis that appear in my SAS dataset. May you please help me with an example? Thanks

[Ksharp]

You could list all the characters you want keep as you did with 'k'.

new_var=compress(var,"ABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890$%.-'"":|&[]{}*!@#()!/+, ", 'k'
);

[Emma_at_SAS]

Thank you, Ksharp. This is what I was using the 'k' modifier. I was hoping to be able to use the 'g' modifier to only exclude the graphic characters. Maybe at some point, I learn to work efficiently with the modifiers. Thanks!

[Ksharp]

exclude 'g' --> could use 'g' instead of 'k'.
new_var=compress(var,"ABC " ,'g');