difference between two dates by row per customer_id

[Ronein]

Hello

I want to calculate the difference between two dates by row per customer_id.

Please find the raw data

data a;
  input CUSTOMER_ID date anydtdte23.;
  format date date9.;
cards;
1057086 21AUG2014
1057086 25AUG2014
1057086 17SEP2014
1057086 17SEP2014
1057086 19SEP2014
1057086 26SEP2014
9888221 01Feb2014
9888221 23Apr2014
9888221 23Apr2014
9888221 26May2014
9888221 27May2014
;
Run;

[Kurt_Bremser]

Use the dif() function, and set the result to missing if first.customer_id.

[Ronein]

Thank you.

For second customer 9888221 in first row I get -237 instead of null.

Why?

data a;
  input CUSTOMER_ID date anydtdte23.;
  format date date9.;
cards;
1057086 21AUG2014
1057086 25AUG2014
1057086 17SEP2014
1057086 17SEP2014
1057086 19SEP2014
1057086 26SEP2014
9888221 01Feb2014
9888221 23Apr2014
9888221 23Apr2014
9888221 26May2014
9888221 27May2014
;
Run;

proc sort data=a; by customer_id date; run;

Data b;
set a;
days_between = dif(date);
IF first.customer_id then days_between=.; 
Run;

[Ksharp]

Data b;
set a;


by customer_id ;


days_between = dif(date);
IF first.customer_id then days_between=.; 
Run;

[Kurt_Bremser]

first. and last. variables are only set if a proper by statement is present. Using a first. variable without a by causes a related message in the log.