Hi,
Im not quite sure how to approach this question, because the code below works... sometimes. I guess Im trying to understand what the underlying issue\bug(?) is and if there is either a way to catch the error, or alternatively if there is an elegant way to approach this problem.
Anyhoo, here is the code:
DATA d;
LENGTH inputtype $1.
FN $500.
FN2 $250.;
INFILE DATALINES DSD delimiter='|';
INPUT inputtype $ FN $ FN2 $;
SELECT (inputtype);
WHEN ('N') INFILE foo filevar=fn memvar=fn2 LRECL=500 end=done TRUNCOVER;
WHEN ('Z') INFILE foo ZIP filevar=fn memvar=fn2 LRECL=500 end=done TRUNCOVER;
OTHERWISE;
END;
DO while (not done);
INPUT &datain.;
OUTPUT;
END;
DROP inputtype;
DATALINES;
N|N:\...\data|2011.txt
N|N:\...\data|2010.txt
Z|N:\...\dataarchive\2000_2009.zip|2009.txt
Z|N:\...\dataarchive\2000_2009.zip|2008.txt
...
Z|N:\...\dataarchive\2000_2009.zip|2000.txt
;
The problem the code is trying to solve is that my group has annual data stored in fixed width text files. However, at some point the files get archived into ZIP folders. However, since the formatting has not changed I want to pull all the data in to a single data set. So I started looking for a way to change between the normal read and the ZIP read engines.
Again the code above does this. -- Sometimes. But there are a lot of subtle changes, that cuase it to fail, that have no bussiness being an issue.
However, if I switch the WHEN statements around so that it reads:
SELECT (inputtype);
WHEN ('Z') INFILE foo ZIP filevar=fn memvar=fn2 LRECL=500 end=done TRUNCOVER;
WHEN ('N') INFILE foo filevar=fn memvar=fn2 LRECL=500 end=done TRUNCOVER;
OTHERWISE;
END;
I get a ERROR: INVALID RECORD FORMAT on the first line, of the first data set.
Or if I switch the data in DATALINES around so that Zip Files come before the non-ZIP files. I get an INVALID RECORD FORMAT at the start of the non-ZIP lines.
Really if the ZIP device-type on INVALUE in any way comes before the non-ZIPPED version errors get thrown.
The issue persists so that if I take the code that works. Then I run it a second time. I get an ERROR:INVALID RECORD FORMAT. on the first line of the data. (But if I do run it again, then it runs fine).
Additionally,
If I change the placeholder file refernce on one of the infile statements so I have something like:
WHEN ('Z') INFILE foo ZIP filevar=fn memvar=fn2 LRECL=500 end=done TRUNCOVER;
WHEN ('N') INFILE bar filevar=fn memvar=fn2 LRECL=500 end=done TRUNCOVER;
I get the INVALID RECORD FORMAT error.
Thoughts as to whats going on? Is there a way of dynamically change what read engine SAS is using that is a bit less fickle, or at least get around the error that is triggered by a rerun?
Anyhoo if it helps Im running SAS 9.04.01M3
On Enterprise Guide 7.15 HF9
Click image to register for webinar
