Proc Format Invalue regex hierarchy

[sc5]

Hi,

Is there a way to use proc format with regex so that it sequentially evaluates conditions in the order given? 

Example: 

proc format;
	invalue $Fr
	"s/(.*)(PELICA|UNDEREDS|GRASSHOP)(.*)/PELICA/i"	(regexpe) = _same_
	"s/(.*)(DEATHS|DT|CID)(.*)/DEATHS/i"	(regexpe) = _same_
	"s/(.*)(GRA)(.*)/LLL/i"	(regexpe) = _same_
	"s/(.*)(EMPRE)(.*)/EMPRESSS/i"	(regexpe) = _same_
	other = ''
	;
quit;
data fruit;
	length fruit $50. ;
	infile cards dsd dlm='>' truncover;
	input fruit ;
	cards;
	DEATHS
	PELICA
	PRIZEDE UNDEREDS/EMPRESSS
	GRASSHOP
	GRAMS
run;
data fruit2;
	set fruit;
	fruitset=input(fruit,$Fr.);
run;

The output is 

But I want the output to be: 

i.e. I want the conditions in proc format to be evaluated in the order given. The 3rd row matches the 1st regex condition (the word 'UNDEREDS'), so the output should be PELICA, not EMPRESSS. I only want the output to be EMPRESSS if the conditions before are not met.

Any idea how I can achieve this? Or is there another way I can format strings based on patterns in a specific order of preference? Worst case I'll resort to if-then conditions, but I have a lot of columns to parse with a lot of conditions, and a huge dataset, so I'm trying to do it in a more modular/elegant way. 

Thank you!

[ballardw · Accepted solution]

See if this helps:

proc format;
	invalue $Fr  (notsorted)
	"s/(.*)(PELICA|UNDEREDS|GRASSHOP)(.*)/PELICA/i"	(regexpe) = _same_
	"s/(.*)(DEATHS|DT|CID)(.*)/DEATHS/i"	(regexpe) = _same_
	"s/(.*)(GRA)(.*)/LLL/i"	(regexpe) = _same_
	"s/(.*)(EMPRE)(.*)/EMPRESSS/i"	(regexpe) = _same_
	other = ''
	;
quit;

[sc5]

This seems to work, thanks a lot!

[Oligolas]

Hi,

absolutely. You could also define your formats as a chain with the keyword 'other' to control in detail what happens:

PROC FORMAT;
	invalue $Fr1_ "s/(.*)(PELICA|UNDEREDS|GRASSHOP)(.*)/PELICA/i"	(regexpe) = _same_ other=[$Fr2_.];
	invalue $Fr2_ "s/(.*)(DEATHS|DT|CID)(.*)/DEATHS/i"	(regexpe) = _same_ other=[$Fr3_.];
	invalue $Fr3_ "s/(.*)(GRA)(.*)/LLL/i"	(regexpe) = _same_ other=[$Fr4_.];
	invalue $Fr4_ "s/(.*)(EMPRE)(.*)/EMPRESSS/i"	(regexpe) = _same_ other=' ';
	;
QUIT;

[sc5]

Ooh ok I didn't know we could do that, thank you!