| Date: | Tue, 2 Oct 2001 04:19:23 -0700 |
|---|
| Reply-To: | Janis Jekabsons <Janis.Jekabsons@DRESDNER-BANK.COM> |
|---|
| Sender: | "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU> |
|---|
| From: | Janis Jekabsons <Janis.Jekabsons@DRESDNER-BANK.COM> |
|---|
| Organization: | http://groups.google.com/ |
|---|
| Subject: | Reading a .csv with commas in fields |
|---|
| Content-Type: | text/plain; charset=ISO-8859-1 |
|---|
Hi all,
I am reading a comma-separated flat file. Some fields may contain
a comma preceded by the \ escape character. So the data looks like this:
12345,One Way Drive 45\, 15,6789
which should be read like
ONE=12345 TWO=One Way Drive 45, 15 THREE=6789
To get around the problem I use the conditional input -- keeping
the input line in buffer, then checking if the last character
of TWO is a backslash and acting accordingly. So the program looks like this,
libname lib "/blabla";
filename da_d1 "/blabla/bla.txt";
data lib.da_d1;
infile da_d1 dsd dlm=',' missover lrecl=590 recfm=v;
input bla1 : $19.
@
;
if substr(bla1,length(bla1),1)='\' then do;
input
bla1_ : $27.
bla2 : $27.
/* And now check bla2 also ... */
;
bla1 = left(trim(substr(bla1,1,length(bla1)-1)))||','||left(trim(bla1_));
end;
else do;
input
bla2 : $27.
/* And now check bla2 also ... */
;
end;
run;
The problem is, of course, that, if this occurs in many fields,
the program becomes unwieldy because of the nested do/else.
Maybe some kind soul out there knows of and can share a clue
that can make this simpler? I am using SAS 8.1 on UNIX.
Regards,
Janis Jekabsons
| 
