topic Converting Date columns into Char. in SAS Programming

Converting Date columns into Char.

kuridisanjeev — Mon, 01 Oct 2012 09:08:23 GMT

Hi...

I am having the following code

data xx;

input folder $ date ddmmyy10.;

format date ddmmyy10.;

cards;

week1 21/05/2012

week2 21/05/2011

week3 21/05/2010

;

run;

data yy;

set xx;

format week1 week2 week3 date9.;

if folder = 'week1' then week1 = date;

if folder = 'week2' then week2 = date;

if folder = 'week3' then week3 = date;

run;

I want to display ' '(space) in the place of ' . '(period) , whenever the date is missing the Week1,week2,week3 columns .

Note:I don't want to use any IF conditions.

Thanks in advance ..

Regards,

Sanjeev.K

Re: Converting Date columns into Char.

Amir — Mon, 01 Oct 2012 09:31:11 GMT

Hi Sanjeev,

You could try using the following options statement before your code:

options missing=' ';

I wasn't sure if you wanted to replace your if statements as you noted, but if that is the case then you could use the select statement.

Regards,

Amir.

Re: Converting Date columns into Char.

kuridisanjeev — Mon, 01 Oct 2012 09:43:17 GMT

Hi Amir..

IF i use Missing='' option,other numeric variables(which are not shown in the code) also effected.but i want to keep periods in numeric variables when there is no data.

Thanks..

Sanjeev.K

Re: Converting Date columns into Char.

Amir — Mon, 01 Oct 2012 11:30:24 GMT

Hi Sanjeev,

I'm not sure how some numeric data can have missing represented by a blank and others be represented by a period at the same time

Perhaps consider using some default value which is not going to appear in your input data.

I did find this forum question https://communities.sas.com/message/139151#139151 but was unable to get the solution to fit around your problem.

Regards,

Amir.

Message was edited by: Amir Malik - rephrase.

Re: Converting Date columns into Char.

LinusH — Mon, 01 Oct 2012 12:01:12 GMT

To avoid your if-statements, use PROC TRANSPOSE.

I think as Amir it's hard to have different missing chars in he same report.

perhaps (untested!) you can create a new format based on ddmmyy, where you explicitly set the missing to space...?

Re: Converting Date columns into Char.

LinusH — Mon, 01 Oct 2012 12:32:40 GMT

Another easier option is to convert your date columns to char prior to reporting, and then replace . with space.

Re: Converting Date columns into Char.

yaswanthj — Mon, 01 Oct 2012 12:35:56 GMT

Could you please provide what option is there.

Thanks & Regards,

Yaswanth J.

Re: Converting Date columns into Char.

Patrick — Mon, 01 Oct 2012 12:36:32 GMT

You could create your own format based on an existing one like done in below code:

proc format;
value My_ddmmyy
.=' '
other=[ddmmyy.]
;
run;

data have;
input folder $ date ddmmyy10.;
format date My_ddmmyy10.;
cards;
week1 21/05/2012
week2
week3 21/05/2010
;
run;

Re: Converting Date columns into Char.

kuridisanjeev — Mon, 01 Oct 2012 12:37:26 GMT

Hi LinusH..

Did't get you completely...

Can you please write sample code for my above example.??

That make sense..

Thanks..

Sanjeev.K

Re: Converting Date columns into Char.

LinusH — Mon, 01 Oct 2012 12:45:46 GMT

If you report directly from your import data, use char informat directly:

input $date char10.;

The there will never be a period (unless your source data contains it...)

PROC TRANSPOSE, see online doc, stuffed with examples...

Re: Converting Date columns into Char.

Patrick — Mon, 01 Oct 2012 12:51:02 GMT

If there is not a very good reason then I wouldn't store dates as strings but as SAS date values. If you're using strings then one of the issues you're facing in reports is the sort order of the dates in your reports.

I believe using a user defined format as posted earlier will cover the requirement without causing new issues.

Re: Converting Date columns into Char.

ballardw — Mon, 01 Oct 2012 14:54:56 GMT

Patrick's response is a very good idea in general. When only one or two values need special appearance make a custom version of the format like this. It can save a great deal of complex coding.