Date: Tue, 30 Apr 2002 10:57:43 +0100
Reply-To: Peter Crawford <peter.crawford@DB.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Peter Crawford <peter.crawford@DB.COM>
Subject: Re: Storing Infile error messages in sas dataset
Content-type: text/plain; charset=iso-8859-1
I don't know any short-cut to a sas dataset with the input errors,
but I do something like...
To capture this info, I re-route the log with proc printto.
Then during the input data step, pay attention to _error_.
if _error_ then
do;
put 'errors surfaced on ' row= ;
_error_ =0;
_err ++1 ; retain _err 0;
/* retain statement only foe emphasis */
end;
row +1;
return;
If at any time _error_ becomes non-zero, then the step
will (at end of data), call execute() to email that log to
those who have to resolve the data errors.
The data step uses the column titles of the input
file, as variable names, so the columns in error are
easily recognised by a business user.
By switching off _error_ (after flagging errors
are present) the log is not filled with the put _all_
messages and _infile_ . To locate the data causing
_error_, the line number and column name (in error)
are enough, like:
NOTE: Invalid data for B in line 12 3-3.
Column position is irrelevant because the data is .csv
but easy to ignore (g)
If this is the info to be stored in a dataset of infile errors,
it should be simple to parse.
Hope you find the design useful
Regards
Peter Crawford
Datum: 30/04/2002 02:25
An: SAS-L@LISTSERV.UGA.EDU
Antwort an: Jack Hamilton <JackHamilton@FIRSTHEALTH.COM>
Betreff: Re: Storing Infile error messages in sas dataset
Nachrichtentext:
I'm not sure what you mean. Do you mean that if you have a SAS log that
looks like this:
-----
8 data test;
9 infile cards;
10 input a b c;
11 cards;
NOTE: Invalid data for B in line 12 3-3.
RULE:----+----1----+----2----+----3----+----4----+-
12 1 b 3
A=1 B=. C=3 _ERROR_=1 _N_=1
NOTE: Invalid data for A in line 13 1-1.
NOTE: Invalid data for C in line 13 5-5.
13 a 2 c
A=. B=2 C=. _ERROR_=1 _N_=2
-----
You want to capture the fact that B is invalid in the first input line,
and A & C are invalid in the second line, along with the locations from
which SAS tried to read the values?
If so, I don't know of a straightforward way to do exactly what you
want. You could read the input file one field at a time, checking after
each read to see if there was a conversion error (a missing value
usually indicates a conversion error; in some cases you might have to
use the _ERROR_ automatic variable), and using the COLUMN= option on the
INFILE statement to track the current location in the input line.
But that could get complicated. If you can switch to formatted or
column input mode you won't have to worry about determining the input
position, and your task will be easier.
Why do you want to store this information? Why not just look at the
SAS log?
--
JackHamilton@FirstHealth.com
Manager, Technical Development
METRICS Department, First Health
West Sacramento, California USA
>>> StatDept <StatDept@YAHOO.COM> 04/29/2002 5:07 PM >>>
Hi,
I want to store the field name in the infile error message in a
sasdataset to be printed at a later date. Any ideas?
-SH
--
Diese E-Mail enthält vertrauliche und/oder rechtlich geschützte Informationen. Wenn Sie nicht der richtige Adressat sind oder diese E-Mail irrtümlich erhalten haben, informieren Sie bitte sofort den Absender und vernichten Sie diese Mail. Das unerlaubte Kopieren sowie die unbefugte Weitergabe dieser Mail ist nicht gestattet.
This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and destroy this e-mail. Any unauthorized copying, disclosure or distribution of the material in this e-mail is strictly forbidden.
