topic Re: Macro to loop thru all date variables and convert to character type in SAS Programming

Macro to loop thru all date variables and convert to character type

SAS_Ryan — Tue, 14 Aug 2018 19:46:21 GMT

Is there a way to programmatically loop thru all variables in a dataset and convert any date variables into character variables? I need to push my data to a 3rd party visualization software and it does not like the SAS formatted dates.

This code does the conversion I need, can it be macrotized to automatically convert for every date variable?

data my_dataset;
set my_dataset;
if my_date ne . then
my_date_ch = put(my_date, mmddyy10.);
else
my_date_ch = '';
run;

Re: Macro to loop thru all date variables and convert to character type

PaigeMiller — Tue, 14 Aug 2018 19:49:09 GMT

There's a simpler way than a loop.

Just assign the proper format to the date variables, and then when the data is exported somehow, the date variable appear in the selected format.

Re: Macro to loop thru all date variables and convert to character type

Reeza — Tue, 14 Aug 2018 19:50:07 GMT

An array with a loop is likely all you need....if you have a naming convention this is much easier as well.

data want;
set have;

array _in(*) list of date variables here;
array _out(*) list of new variables here;

do i=1 to dim(_in);
_out = put(_in(i), ddmmyy10.);
end;

run;

@SAS_Ryan wrote:

Is there a way to programmatically loop thru all variables in a dataset and convert any date variables into character variables? I need to push my data to a 3rd party visualization software and it does not like the SAS formatted dates.

This code does the conversion I need, can it be macrotized to automatically convert for every date variable?

data my_dataset;
set my_dataset;
if my_date ne . then
my_date_ch = put(my_date, mmddyy10.);
else
my_date_ch = '';
run;

Re: Macro to loop thru all date variables and convert to character type

ballardw — Tue, 14 Aug 2018 19:57:46 GMT

How many date variables are you talking about? It is likely easier to use a pair of arrays.

Something like:

data want;
   set my_dataset;
   array d my_date <list other date variables>;
   array c $ 10 my_date_ch <list the new names here>;
   do _i_ = 1 to dim(d);
      if d[i] then c[_i_] = put(d[_i_],mmddyy10.);
   end;
run;

Make sure you have the same number of items in both arrays.

if you set

options missing=' ';

you wouldn't even need the test for if the existing date variable isn't missing.

Do they all have the same format?