/******************************************************************************
 * Programme : PROC DATAMETRICS: Erweiterte Datenqualitätsanalyse
 * Reference : PROCDA74D2
 * Source    : https://www.wearecas.eu/en/sampleCode/PROCDA74D2
 ******************************************************************************/

/* --- BLOC 1 --- */
data work.my_data;
    length name $30 address $50 city $20 state $2;
    input name $ address $ city $ state $;
    datalines;
"John Doe" "123 Main St" "Anytown" "NY"
"Jane Smith" "456 Oak Ave" "Anycity" "CA"
"John Doe" "123 Main St" "Anytown" "NY"
"Peter Jones" "789 Pine Ln" "Otherville" "TX"
"Alice Brown" "101 Maple Dr" "Anytown" "NY"
"Bob White" "202 Elm St" "Otherville" "TX"
"Charlie Green" "303 Cedar Rd" "Anycity" "CA"
"David Black" "404 Birch Ct" "Anytown" "NY"
;
run;

proc datametrics data=work.my_data out=work.basic_metrics;
   variables name address;
   run;

proc print data=work.basic_metrics;
   title "Grundlegende Metriken für Name und Adresse";
run;

/* --- BLOC 2 --- */
/* Assurez-vous que work.my_data est déjà créé à partir de l'Exemple 1 */

proc datametrics data=work.my_data out=work.common_metrics frequencies=10 
   minmax=5 median;
   identities qkb='/sas/dqc/QKBLoc' locale='ENUSA' def='Field Content';
   variables name address city;
   run;

proc print data=work.common_metrics;
   title "Metriken mit Häufigkeiten, Min/Max, Median und QKB";
run;

/* --- BLOC 3 --- */
/* Assurez-vous que work.my_data est déjà créé à partir de l'Exemple 1 */

proc format;
   value $statefmt
   'NY'='New York'
   'CA'='California'
   'TX'='Texas'
   other='Autre';
run;

data work.formatted_data;
   set work.my_data;
   format state $statefmt.;
run;

proc datametrics data=work.formatted_data out=work.advanced_metrics 
   frequencies=20 minmax=10 threads=4 format;
   identities qkb='/sas/dqc/QKBLoc' locale='ENUSA' 
   def='Field Content' multiidentity;
   variables name address city state;
   run;

proc print data=work.advanced_metrics;
   title "Erweiterte Metriken mit Formaten, Threads und Multi-Identitäten";
run;

/* --- BLOC 4 --- */
/* Assurez-vous que work.my_data est déjà créé à partir de l'Exemple 1 */

caslib _all_ assign;

proc casutil;
   load data=work.my_data outcaslib='casuser' casout='my_cas_data' replace;
run;

proc datametrics data=casuser.my_cas_data out=casuser.cas_metrics 
   frequencies=5 minmax=3 threads=2;
   identities qkb='/sas/dqc/QKBLoc' locale='ENUSA' def='Field Content';
   variables name address city;
   run;

proc print data=casuser.cas_metrics;
   title "Metriken über DATAMETRICS auf CAS";
run;