|
The nesting of informats to deal with special values (whether retained or not)
works a treat !
As for the use of a quoted string range for a numeric informat, remember that an
informat expects to read text
paste this into your session
proc format fmtlib;
invalue dbdate
'01/01/0001' = .D
other = (| ddmmyy10. |)
;
data xx;
length text $12 ;
attrib db2dat format = date9. informat = dbdate.;
input text $ @1 db2dat ;
if db2dat = .D then
do;
thatmiss = 'null 1' ;
db2dat = . ;
end;
put _all_;
cards;
01/02/99
01/01/01
01/01/0001
;
I got a log like this (winNT 6.12 ts060)
429 cards;
TEXT=01/02/99 DB2DAT=01FEB1999 THATMISS= _ERROR_=0 _N_=1
TEXT=01/01/01 DB2DAT=01JAN2001 THATMISS= _ERROR_=0 _N_=2
TEXT=01/01/0001 DB2DAT=. THATMISS=null 1 _ERROR_=0 _N_=3
NOTE: The data set WORK.XX has 3 observations and 3 variables.
NOTE: The DATA statement used 0.18 seconds.
Datum: 02.11.99 21:12
An: Peter Crawford/Zentrale/DeuBaExt
Kopie: sas-l@uga.cc.uga.edu
Betreff: RE: SAS Date problem
Nachrichtentext:
Will that work? I am (slightly) aware of nesting informats, but I did not
realize you could mix character comparisons and numeric informats like that.
If so, then the original poster can either use your informat as supplied
(although my browser ruins the square brackets around the nested informat)
or change the .D to plain . and keep me happy.
Tim Berryhill - Contract Programmer and General Wizard
TWB2@PGE.COM or http://www.aartwolf.com/twb.html
Frequently at Pacific Gas & Electric Co., San Francisco
The correlation coefficient between their views and
my postings is slightly less than 0
> ----------
> From: peter.crawford@db.com[SMTP:peter.crawford@db.com]
> Sent: Tuesday, November 02, 1999 12:05 PM
> To: TWB2@pge.com
> Cc: sas-l@uga.cc.uga.edu
> Subject: RE: SAS Date problem
>
>
>
> well that is emphatic enough ... no
> I was considering a user informat to handle such situations, almost
> transparently
> invalue dbdate
> '01/01/0001' = .D
> other = (! ddmmyy10. !)
> ;
> I don't like to throw away the info, so to meet Tim's preference, I'd
> follow the
> use of such an informat with data step statements like
> if thatdate = .D then
> do;
> thatmiss = 'null 1' ;
> thatdate = . ;
> end;
>
> Clearly, this isn't how the db2 interface would work, but.....
>
>
>
>
> Datum: 02.11.99 20:52
> An: Peter Crawford/Zentrale/DeuBaExt
> sas-l@uga.cc.uga.edu
>
>
>
> Betreff: RE: SAS Date problem
> Nachrichtentext:
>
>
>
> Peter, Perhaps a special missing value is appropriate here, perhaps
> standard
> missing values are appropriate here. I tend to avoid the special missing
> values. Casual coders test for equality to . more often than they test
> for
> less than or equal to .Z, so unless there is a specific reason for the
> missing value, I just use .. Even if there is a specific reason for the
> missing value, I just use .. If there are a variety of specific reasons
> for
> missing values, I still just use ., and I add a field for storing the
> reason.
> Tim
> > ----------
> > From: peter.crawford@db.com[SMTP:peter.crawford@db.com]
> > Sent: Tuesday, November 02, 1999 11:19 AM
> > To: TWB2@pge.com
> > Cc: SAS-L@listserv.uga.edu
> > Subject: Re: SAS Date problem
> >
> > would you consider using a special missing like .D for specifics like
> this
> > ?
> >
> <SNIP>
>
>
>
>
| 
