Witch function I can to use for search a date inside chars: PRXCHANGE, PRXMATCH, PRXPARSE, another?

[Mozer19]

Witch function I can to use for search a date inside chars: PRXCHANGE, PRXMATCH, PRXPARSE, another one?

Data example;

char = "this_is_a_example_with_date_20230331";

/* The format date is YYYYMMDD */

resolve = ? / * function;

run;

output:

resolve

31/03/2023

[PGStats · Accepted solution]

PRXMATCH

data test;
char = "this_is_a_example_with_date_20230331";
pos = prxmatch("/(19|20)\d\d[01]\d[0123]\d/o", char);
if pos > 0 then validDate = input(substr(char,pos,8),yymmdd8.);
format validDate yymmdd10.;
drop pos;
run;

proc print;

[Mozer19]

@PGStats , Thank you so much! Your solution solved my problem.

[Mozer19]

Hi, @PGStats. Are you ok?
I try to learn your solution's about prxmatch, what does that means (19|20) inside the function? Exemple: pos = prxmatch("/(19|20)\d\d[01]\d[0123]\d/o", char);

If you explain me more about every single steps of function, like a, how the function transform 20230406 in 06-04-2023, I preciate this. Can you?


Would You to show me more a function to return month and year?
And another function to return only year?

[PGStats]

Here is a commented version of the code:

data test;
char = "this_is_a_example_with_date_20230331";
/*
The first parameter of the prxmatch function is a 
Perl regular expression. It is interpreted as follows
Find a substring that:
starts with 19 or 20, followed by two digits, followed by 0 or 1,
followed by a digit, followed by 0, 1, 2, or 3, followed by a digit.
If found, return the position of the beginning the matching substring.
*/
pos = prxmatch("/(19|20)\d\d[01]\d[0123]\d/o", char);
/*
If a match is found, extract the substring, read it as a SAS date with 
informat YYMMDD8.
*/
if pos > 0 then validDate = input(substr(char,pos,8),yymmdd8.);
/* 
When displaying the SAS date, use the format YYMMDD10.
*/
format validDate yymmdd10.;

drop pos;
run;

Perl regular expression syntax is described here.

To display the date as year and month, use format YYMMD.

To display the date as year only, use format YEAR4.

Note that the period is part of the format name.

hth