Weekly Data Read

[pangea17]

I have two years' worth of daily data. I want to break it up into weekly data Monday to Sunday.  I was going to write 104 if/then/else statements like the following, but I decided to ask if there is a better way.  I am using enterprise guide 7.1 and sas 9.4.  I am just comparing various metrics from week to week.

else if 'Pres Received Date'n gt '30dec2018'd and 'Pres Received Date'n le '06jan2019'd then seg = 1;

else if 'Pres Received Date'n gt '06jan2019'd and 'Pres Received Date'n le '13jan2019'd then seg = 2;

[PeterClemmensen · Accepted solution]

Definitely do not write 104 if-then statements 🙂

Is it important that seg is 1,2,3 and so on? If not, check out the WeekU. Format that will bin the variable for you

[Reeza]

Use the WEEK() function? There are several variations that can be customzied for what you need.

Depending on how you're summarizing, you may be able to apply a format instead of creating a new variable as well.

https://documentation.sas.com/?docsetId=lefunctionsref&docsetTarget=n1ka2ulrvrjlasn0z7beco2yrgas.htm...

This is an example of using a formatted date in a summary table. Although dates are not annual, the summary is annual.

proc means data=sashelp.stocks N NMISS min max mean median STACKODS;
class date;
format date year4.;
var open high low;
output out=want n= mean= sum= median= mean= / autoname autolabel;
ods output summary = want2;
run;

---

@pangea17 wrote:

I have two years' worth of daily data. I want to break it up into weekly data Monday to Sunday.  I was going to write 104 if/then/else statements like the following, but I decided to ask if there is a better way.  I am using enterprise guide 7.1 and sas 9.4.  I am just comparing various metrics from week to week.

else if 'Pres Received Date'n gt '30dec2018'd and 'Pres Received Date'n le '06jan2019'd then seg = 1;

else if 'Pres Received Date'n gt '06jan2019'd and 'Pres Received Date'n le '13jan2019'd then seg = 2;

---

[ballardw]

---

@pangea17 wrote:

I have two years' worth of daily data. I want to break it up into weekly data Monday to Sunday.  I was going to write 104 if/then/else statements like the following, but I decided to ask if there is a better way.  I am using enterprise guide 7.1 and sas 9.4.  I am just comparing various metrics from week to week.

else if 'Pres Received Date'n gt '30dec2018'd and 'Pres Received Date'n le '06jan2019'd then seg = 1;

else if 'Pres Received Date'n gt '06jan2019'd and 'Pres Received Date'n le '13jan2019'd then seg = 2;

---

Unless you have some shop requirement I would modify anything that requires name literals like: 'Pres Received Date'n 

It might help to provide a small sample of your data and what the desired result would be, and if the result is a report (people read) or a data set needed by another process.

Instructions here: https://communities.sas.com/t5/SAS-Communities-Library/How-to-create-a-data-step-version-of-your-dat... will show how to turn an existing SAS data set into data step code that can be pasted into a forum code box using the {i} icon or attached as text to show exactly what you have and that we can test code against.

A format is very likely the way to go. I don't like the default behaviors of the SAS week formats:

data example;
  do d='25DEC2018'd to '18FEB2020'd;
   output;
  end;
run;
proc format library=work;
/* %0W and %0U use different defintions for 
   start of week and when to count a complete
   week the %0V
*/
picture myweek  (default=8)
low-high = '%Y-%0V'   (datatype=date)
;
run;

proc freq data=example;
   format d  myweek.;
run;

Use V, U or W , check the documentation on either the Picture statement in Proc Format or the WEEK function with the U, V and W descriptor.